You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openjpa.apache.org by David Ezzio <de...@apache.org> on 2008/02/01 01:08:01 UTC
Failures in TeamCity Derby build and svn commit: r616972
Hi,
Progress has been slow on the test failures apparently introduced by
616972. I have made progress in reproducing the results.
The miscounting in TeamCity is gross. Although it lists 33 failing
tests complete with a credible stack trace for each, in fact, there is
only one failing test.
Meanwhile, I believe now, that change 616658 is no longer a suspect. I
get very erratic results here, but they seem to vary more by whether I
run "mvn test" after "mvn clean" or "mvn install" after "mvn clean". By
doing just what TeamCity does, "mvn clean compile test", I can finally
reproduce the results locally that TeamCity reports in its Compile-Derby
build for this change (395, I think). In spite of this, if I run "mvn
-Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the test
passes.
It would be handy if I knew a few things. How to pass in VM debug args
to Maven for its test runs, where to set the OpenJPA logging when Maven
runs the tests, and how to pass in the appropriate parameters when
running the test case outside of Maven. Any insights on these questions
will be appreciated.
But at least at this point, I can reproduce locally what TeamCity
reports, and although it happened only once out of one try, I am
optimistic that it is reproducible. I will continue working this
tomorrow.
If the hangup is intolerable, let me know, and I'll back out my change
tomorrow.
David Ezzio
David Ezzio wrote:
> Hi,
>
> I am convinced that change 616658 (small change to DBDictionary.java) is
> the source of the two intermittent failures that I am seeing locally.
> One of those intermittent failures is the failure in
> TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans which is
> one of the 33 related failures currently seen in TeamCity Derby builds.
>
> The other intermittent failure is TestVersion.testVersionTimestamp which
> I reported earlier as occurring locally with change 616658.
>
> When I revert change 616658 locally, both of these issues go away even
> with the changes that I submitted to AbstractBrokerFactory in change
> 616972.
>
> I have a question: is it possible to run a build in TeamCity without
> submitting the change (or conditionally submitting the change) to SVN?
> I'd like to find out if reverting 616658 will fix the issue. If so,
> then I can try to figure out why it's a problem.
>
> Thanks,
>
> David
>
> David Ezzio wrote:
>> No, but I'm taking a look since my changes appear to be the cause.
>>
>> Patrick Linskey wrote:
>>> 5. The tests are now failing in the automated build; any idea why
>>> this might be?
>>>
>>> -Patrick
>>>
>>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com> wrote:
>>>> Hi,
>>>>
>>>> A few questions:
>>>>
>>>> 1. Were you seeing this cause problems somewhere? I.e., when do we
>>>> serialize BrokerFactory instances?
>>>>
>>>> 2. Why not just move the initialization code to the declaration line
>>>> (instead of the constructor)?
>>>>
>>>> 3. It looks like there are other transactional structures that might
>>>> not be properly handled in serialization / deserialization (lifecycle
>>>> listeners, for example). Should we be making more changes there?
>>>>
>>>> -Patrick
>>>>
>>>>
>>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
>>>>> Author: dezzio
>>>>> Date: Wed Jan 30 16:59:02 2008
>>>>> New Revision: 616972
>>>>>
>>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
>>>>> Log:
>>>>> Allow EntityManagerFactory objects to be serialized and
>>>>> deserialized successfully.
>>>>>
>>>>> Added:
>>>>>
>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
>>>>> (with props)
>>>>> Modified:
>>>>>
>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
>>>>>
>>>>>
>>>>> Modified:
>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
>>>>>
>>>>> URL:
>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java?rev=616972&r1=616971&r2=616972&view=diff
>>>>>
>>>>> ==============================================================================
>>>>>
>>>>> ---
>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
>>>>> (original)
>>>>> +++
>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
>>>>> Wed Jan 30 16:59:02 2008
>>>>> @@ -147,8 +147,7 @@
>>>>> */
>>>>> protected AbstractBrokerFactory(OpenJPAConfiguration config) {
>>>>> _conf = config;
>>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>> - ConcurrentReferenceHashSet.WEAK);
>>>>> + getPcClassLoaders();
>>>>> }
>>>>>
>>>>> /**
>>>>> @@ -287,13 +286,13 @@
>>>>> if (needsSub(cls))
>>>>> toRedefine.add(cls);
>>>>> }
>>>>> - _pcClassLoaders.add(loader);
>>>>> + getPcClassLoaders().add(loader);
>>>>> _pcClassNames = c;
>>>>> }
>>>>> _persistentTypesLoaded = true;
>>>>> } else {
>>>>> // reload with this loader
>>>>> - if (_pcClassLoaders.add(loader)) {
>>>>> + if (getPcClassLoaders().add(loader)) {
>>>>> for (Iterator itr = _pcClassNames.iterator();
>>>>> itr.hasNext();) {
>>>>> try {
>>>>> Class cls =
>>>>> @@ -818,4 +817,15 @@
>>>>> _transactional.remove (_trans);
>>>>> }
>>>>> }
>>>>> +
>>>>> + /**
>>>>> + * Method insures that deserialized EMF has this reference
>>>>> re-instantiated
>>>>> + */
>>>>> + private Collection getPcClassLoaders() {
>>>>> + if (_pcClassLoaders == null)
>>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>> + ConcurrentReferenceHashSet.WEAK);
>>>>> +
>>>>> + return _pcClassLoaders;
>>>>> + }
>>>>> }
>>>>>
>>>>> Added:
>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>
>>>>> URL:
>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java?rev=616972&view=auto
>>>>>
>>>>> ==============================================================================
>>>>>
>>>>> ---
>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
>>>>> (added)
>>>>> +++
>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
>>>>> Wed Jan 30 16:59:02 2008
>>>>> @@ -0,0 +1,74 @@
>>>>> +/*
>>>>> + * Licensed to the Apache Software Foundation (ASF) under one
>>>>> + * or more contributor license agreements. See the NOTICE file
>>>>> + * distributed with this work for additional information
>>>>> + * regarding copyright ownership. The ASF licenses this file
>>>>> + * to you 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 org.apache.openjpa.persistence.simple;
>>>>> +
>>>>> +import java.io.*;
>>>>> +
>>>>> +import javax.persistence.EntityManager;
>>>>> +import javax.persistence.EntityManagerFactory;
>>>>> +import javax.persistence.EntityTransaction;
>>>>> +
>>>>> +import junit.textui.TestRunner;
>>>>> +import org.apache.openjpa.persistence.OpenJPAEntityManager;
>>>>> +import org.apache.openjpa.persistence.test.SingleEMFTestCase;
>>>>> +
>>>>> +/**
>>>>> + * Tests that a EntityManagerFactory can be used after serialization.
>>>>> + *
>>>>> + * @author David Ezzio
>>>>> + */
>>>>> +public class TestSerializedFactory
>>>>> + extends SingleEMFTestCase {
>>>>> +
>>>>> + public void setUp() {
>>>>> + setUp(AllFieldTypes.class);
>>>>> + }
>>>>> +
>>>>> + public void testSerializedEntityManagerFactory() throws
>>>>> Exception {
>>>>> + // serialize and deserialize the entity manager factory
>>>>> + ByteArrayOutputStream baos = new ByteArrayOutputStream();
>>>>> + ObjectOutputStream oos = new ObjectOutputStream(baos);
>>>>> + oos.writeObject(emf);
>>>>> + EntityManagerFactory emf2 =
>>>>> + (EntityManagerFactory) new ObjectInputStream(
>>>>> + new
>>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
>>>>> +
>>>>> + // use the deserialized entity manager factory
>>>>> + assertTrue("The deserialized entity manager factory is not
>>>>> open",
>>>>> + emf2.isOpen());
>>>>> + EntityManager em = emf2.createEntityManager();
>>>>> + assertTrue("The newly created entity manager is not open",
>>>>> em.isOpen());
>>>>> +
>>>>> + // exercise the entity manager produced from the
>>>>> deserialized EMF
>>>>> + em.getTransaction().begin();
>>>>> + em.persist(new AllFieldTypes());
>>>>> + em.getTransaction().commit();
>>>>> +
>>>>> + // close the extra resources
>>>>> + em.close();
>>>>> + assertFalse("The entity manager is not closed", em.isOpen());
>>>>> + emf2.close();
>>>>> + assertFalse("The entity manager factory is not closed",
>>>>> emf2.isOpen());
>>>>> + }
>>>>> +
>>>>> + public static void main(String[] args) {
>>>>> + TestRunner.run(TestSerializedFactory.class);
>>>>> + }
>>>>> +}
>>>>> +
>>>>>
>>>>> Propchange:
>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>>
>>>>> svn:eol-style = native
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Patrick Linskey
>>>> 202 669 5907
>>>>
>>>
>>>
>>>
>>
>
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by Kevin Sutter <kw...@gmail.com>.
On Jan 31, 2008 7:56 PM, Patrick Linskey <pl...@gmail.com> wrote:
> To be honest, I don't really understand why a BrokerFactory needs to
> be serializable, so I think probably it'd be good to make sure that
> there is good reason for that before going too deep into this
> debugging exercise.
I tend to agree with this statement. What is the scenario that requires the
EMF (BrokerFactory) to be Serializable? Thanks.
Kevin
>
>
> -Patrick
>
> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
> > > If the hangup is intolerable, let me know, and I'll back out my change
> > > tomorrow.
> >
> > Personally, I'd rather have the TeamCity build be clean while you
> > investigate. As you noted, the error-reporting from surefire is
> > frustratingly inaccurate, and this really erodes the benefit of
> > continuous integration.
> >
> > > It would be handy if I knew a few things. How to pass in VM debug
> args
> > > to Maven for its test runs, where to set the OpenJPA logging when
> Maven
> > > runs the tests, and how to pass in the appropriate parameters when
> > > running the test case outside of Maven. Any insights on these
> questions
> > > will be appreciated.
> >
> > I usually just add log configuration to the test cases themselves, in
> > the setUp(...) method:
> >
> > setUp(Class.class, Class2.class, ..., "openjpa.Log", "SQL=TRACE");
> >
> > I believe that it is also possible to set this at the command line via
> > -Dopenjpa.Log=SQL=TRACE types of clauses.
> >
> > -Patrick
> >
> >
> > On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
> > > Hi,
> > >
> > > Progress has been slow on the test failures apparently introduced by
> > > 616972. I have made progress in reproducing the results.
> > >
> > > The miscounting in TeamCity is gross. Although it lists 33 failing
> > > tests complete with a credible stack trace for each, in fact, there is
> > > only one failing test.
> > >
> > > Meanwhile, I believe now, that change 616658 is no longer a suspect.
> I
> > > get very erratic results here, but they seem to vary more by whether I
> > > run "mvn test" after "mvn clean" or "mvn install" after "mvn clean".
> By
> > > doing just what TeamCity does, "mvn clean compile test", I can finally
> > > reproduce the results locally that TeamCity reports in its
> Compile-Derby
> > > build for this change (395, I think). In spite of this, if I run "mvn
> > > -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the test
> > > passes.
> > >
> > > It would be handy if I knew a few things. How to pass in VM debug
> args
> > > to Maven for its test runs, where to set the OpenJPA logging when
> Maven
> > > runs the tests, and how to pass in the appropriate parameters when
> > > running the test case outside of Maven. Any insights on these
> questions
> > > will be appreciated.
> > >
> > > But at least at this point, I can reproduce locally what TeamCity
> > > reports, and although it happened only once out of one try, I am
> > > optimistic that it is reproducible. I will continue working this
> > > tomorrow.
> > >
> > > If the hangup is intolerable, let me know, and I'll back out my change
> > > tomorrow.
> > >
> > > David Ezzio
> > >
> > >
> > > David Ezzio wrote:
> > > > Hi,
> > > >
> > > > I am convinced that change 616658 (small change to DBDictionary.java)
> is
> > > > the source of the two intermittent failures that I am seeing
> locally.
> > > > One of those intermittent failures is the failure in
> > > > TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans which
> is
> > > > one of the 33 related failures currently seen in TeamCity Derby
> builds.
> > > >
> > > > The other intermittent failure is TestVersion.testVersionTimestampwhich
> > > > I reported earlier as occurring locally with change 616658.
> > > >
> > > > When I revert change 616658 locally, both of these issues go away
> even
> > > > with the changes that I submitted to AbstractBrokerFactory in change
> > > > 616972.
> > > >
> > > > I have a question: is it possible to run a build in TeamCity without
> > > > submitting the change (or conditionally submitting the change) to
> SVN?
> > > > I'd like to find out if reverting 616658 will fix the issue. If so,
> > > > then I can try to figure out why it's a problem.
> > > >
> > > > Thanks,
> > > >
> > > > David
> > > >
> > > > David Ezzio wrote:
> > > >> No, but I'm taking a look since my changes appear to be the cause.
> > > >>
> > > >> Patrick Linskey wrote:
> > > >>> 5. The tests are now failing in the automated build; any idea why
> > > >>> this might be?
> > > >>>
> > > >>> -Patrick
> > > >>>
> > > >>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
> wrote:
> > > >>>> Hi,
> > > >>>>
> > > >>>> A few questions:
> > > >>>>
> > > >>>> 1. Were you seeing this cause problems somewhere? I.e., when do
> we
> > > >>>> serialize BrokerFactory instances?
> > > >>>>
> > > >>>> 2. Why not just move the initialization code to the declaration
> line
> > > >>>> (instead of the constructor)?
> > > >>>>
> > > >>>> 3. It looks like there are other transactional structures that
> might
> > > >>>> not be properly handled in serialization / deserialization
> (lifecycle
> > > >>>> listeners, for example). Should we be making more changes there?
> > > >>>>
> > > >>>> -Patrick
> > > >>>>
> > > >>>>
> > > >>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
> > > >>>>> Author: dezzio
> > > >>>>> Date: Wed Jan 30 16:59:02 2008
> > > >>>>> New Revision: 616972
> > > >>>>>
> > > >>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
> > > >>>>> Log:
> > > >>>>> Allow EntityManagerFactory objects to be serialized and
> > > >>>>> deserialized successfully.
> > > >>>>>
> > > >>>>> Added:
> > > >>>>>
> > > >>>>>
> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> (with props)
> > > >>>>> Modified:
> > > >>>>>
> > > >>>>>
> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > > >>>>>
> > > >>>>>
> > > >>>>> Modified:
> > > >>>>>
> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > > >>>>>
> > > >>>>> URL:
> > > >>>>>
> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java?rev=616972&r1=616971&r2=616972&view=diff
> > > >>>>>
> > > >>>>>
> ==============================================================================
> > > >>>>>
> > > >>>>> ---
> > > >>>>>
> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > > >>>>> (original)
> > > >>>>> +++
> > > >>>>>
> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > > >>>>> Wed Jan 30 16:59:02 2008
> > > >>>>> @@ -147,8 +147,7 @@
> > > >>>>> */
> > > >>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
> config) {
> > > >>>>> _conf = config;
> > > >>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
> > > >>>>> - ConcurrentReferenceHashSet.WEAK);
> > > >>>>> + getPcClassLoaders();
> > > >>>>> }
> > > >>>>>
> > > >>>>> /**
> > > >>>>> @@ -287,13 +286,13 @@
> > > >>>>> if (needsSub(cls))
> > > >>>>> toRedefine.add(cls);
> > > >>>>> }
> > > >>>>> - _pcClassLoaders.add(loader);
> > > >>>>> + getPcClassLoaders().add(loader);
> > > >>>>> _pcClassNames = c;
> > > >>>>> }
> > > >>>>> _persistentTypesLoaded = true;
> > > >>>>> } else {
> > > >>>>> // reload with this loader
> > > >>>>> - if (_pcClassLoaders.add(loader)) {
> > > >>>>> + if (getPcClassLoaders().add(loader)) {
> > > >>>>> for (Iterator itr = _pcClassNames.iterator();
> > > >>>>> itr.hasNext();) {
> > > >>>>> try {
> > > >>>>> Class cls =
> > > >>>>> @@ -818,4 +817,15 @@
> > > >>>>> _transactional.remove (_trans);
> > > >>>>> }
> > > >>>>> }
> > > >>>>> +
> > > >>>>> + /**
> > > >>>>> + * Method insures that deserialized EMF has this reference
> > > >>>>> re-instantiated
> > > >>>>> + */
> > > >>>>> + private Collection getPcClassLoaders() {
> > > >>>>> + if (_pcClassLoaders == null)
> > > >>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
> > > >>>>> + ConcurrentReferenceHashSet.WEAK);
> > > >>>>> +
> > > >>>>> + return _pcClassLoaders;
> > > >>>>> + }
> > > >>>>> }
> > > >>>>>
> > > >>>>> Added:
> > > >>>>>
> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>>
> > > >>>>> URL:
> > > >>>>>
> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java?rev=616972&view=auto
> > > >>>>>
> > > >>>>>
> ==============================================================================
> > > >>>>>
> > > >>>>> ---
> > > >>>>>
> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> (added)
> > > >>>>> +++
> > > >>>>>
> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> Wed Jan 30 16:59:02 2008
> > > >>>>> @@ -0,0 +1,74 @@
> > > >>>>> +/*
> > > >>>>> + * Licensed to the Apache Software Foundation (ASF) under one
> > > >>>>> + * or more contributor license agreements. See the NOTICE file
> > > >>>>> + * distributed with this work for additional information
> > > >>>>> + * regarding copyright ownership. The ASF licenses this file
> > > >>>>> + * to you 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 org.apache.openjpa.persistence.simple;
> > > >>>>> +
> > > >>>>> +import java.io.*;
> > > >>>>> +
> > > >>>>> +import javax.persistence.EntityManager;
> > > >>>>> +import javax.persistence.EntityManagerFactory;
> > > >>>>> +import javax.persistence.EntityTransaction;
> > > >>>>> +
> > > >>>>> +import junit.textui.TestRunner;
> > > >>>>> +import org.apache.openjpa.persistence.OpenJPAEntityManager;
> > > >>>>> +import org.apache.openjpa.persistence.test.SingleEMFTestCase;
> > > >>>>> +
> > > >>>>> +/**
> > > >>>>> + * Tests that a EntityManagerFactory can be used after
> serialization.
> > > >>>>> + *
> > > >>>>> + * @author David Ezzio
> > > >>>>> + */
> > > >>>>> +public class TestSerializedFactory
> > > >>>>> + extends SingleEMFTestCase {
> > > >>>>> +
> > > >>>>> + public void setUp() {
> > > >>>>> + setUp(AllFieldTypes.class);
> > > >>>>> + }
> > > >>>>> +
> > > >>>>> + public void testSerializedEntityManagerFactory() throws
> > > >>>>> Exception {
> > > >>>>> + // serialize and deserialize the entity manager factory
> > > >>>>> + ByteArrayOutputStream baos = new
> ByteArrayOutputStream();
> > > >>>>> + ObjectOutputStream oos = new ObjectOutputStream(baos);
> > > >>>>> + oos.writeObject(emf);
> > > >>>>> + EntityManagerFactory emf2 =
> > > >>>>> + (EntityManagerFactory) new ObjectInputStream(
> > > >>>>> + new
> > > >>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
> > > >>>>> +
> > > >>>>> + // use the deserialized entity manager factory
> > > >>>>> + assertTrue("The deserialized entity manager factory is
> not
> > > >>>>> open",
> > > >>>>> + emf2.isOpen());
> > > >>>>> + EntityManager em = emf2.createEntityManager();
> > > >>>>> + assertTrue("The newly created entity manager is not
> open",
> > > >>>>> em.isOpen());
> > > >>>>> +
> > > >>>>> + // exercise the entity manager produced from the
> > > >>>>> deserialized EMF
> > > >>>>> + em.getTransaction().begin();
> > > >>>>> + em.persist(new AllFieldTypes());
> > > >>>>> + em.getTransaction().commit();
> > > >>>>> +
> > > >>>>> + // close the extra resources
> > > >>>>> + em.close();
> > > >>>>> + assertFalse("The entity manager is not closed",
> em.isOpen());
> > > >>>>> + emf2.close();
> > > >>>>> + assertFalse("The entity manager factory is not closed",
> > > >>>>> emf2.isOpen());
> > > >>>>> + }
> > > >>>>> +
> > > >>>>> + public static void main(String[] args) {
> > > >>>>> + TestRunner.run(TestSerializedFactory.class);
> > > >>>>> + }
> > > >>>>> +}
> > > >>>>> +
> > > >>>>>
> > > >>>>> Propchange:
> > > >>>>>
> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>>
> > > >>>>>
> ------------------------------------------------------------------------------
> > > >>>>>
> > > >>>>> svn:eol-style = native
> > > >>>>>
> > > >>>>>
> > > >>>>>
> > > >>>>
> > > >>>>
> > > >>>> --
> > > >>>> Patrick Linskey
> > > >>>> 202 669 5907
> > > >>>>
> > > >>>
> > > >>>
> > > >>>
> > > >>
> > > >
> > >
> >
> >
> >
> > --
> > Patrick Linskey
> > 202 669 5907
> >
>
>
>
> --
> Patrick Linskey
> 202 669 5907
>
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by Patrick Linskey <pl...@gmail.com>.
Whether or not an EMF needs to be serializable, we still need to
decide if a BrokerFactory needs to be serializable.
For example, currently, you can serialize a Broker, but it will only
deserialize if the VM it's deserializing into has the same
BrokerFactory already deployed.
-Patrick
On Feb 6, 2008 9:19 AM, Craig L Russell <Cr...@sun.com> wrote:
> I think we should ask the JPA EG to resolve whether EMF should be
> serializable.
>
> When a session bean is passivated of saved for failover, it is
> serialized. When serializing the state of a session bean, the fields
> of the bean need to be serialized. If the EMF is stored in a field,
> then it needs to be restored when the session is restored. It would be
> nice if the semantics of passivation and save for failover of an EMF
> were standardized.
>
> Craig
>
>
> On Feb 5, 2008, at 5:01 PM, David Ezzio wrote:
>
> > Hi Patrick,
> >
> > The build was fragile because a couple of test cases were fragile
> > (one did not clean out data before checking whether only the data it
> > added was present, and the other did not ensure that time passed
> > between updates when using a timestamp version check), and because
> > the testing framework itself was fragile in that it attempted to
> > delete data using a new entity manager when transactions might
> > remain open (due to test failures), resulting in deadlock in the
> > database. I've fixed these issues.
> >
> > Since there is no present resolution of the question whether to
> > serialize or not the EMF, I have put in my change once again
> > together with the updated test case. I expect that TeamCity will
> > now be happy due to the aforementioned fixes of the testing
> > framework, etc.
> >
> > Of course, we can always revisit the issue of whether to allow the
> > EMF to be serialized, but since it currently is serializable, I put
> > in the fix to actually allow it to serialize successfully.
> >
> > David
> >
> > Patrick Linskey wrote:
> >> Yes, I agree that if we make serialization not work, we should make
> >> the BF not implement Serializable. That is my preferred approach:
> >> even
> >> if Kodo needs to serialize a PMF for JNDI purposes or something, we
> >> could do that in the PMF impl, instead of in the BF.
> >> IMO, the build is fragile because it runs multiple unrelated tests in
> >> the same VM with no strong isolation between tests. This leads to
> >> unexpected side effects sometimes. This can be frustrating, but the
> >> benefits are considerable: faster test run time, plus we get to find
> >> out about weird unintended side effects.
> >> -Patrick
> >> On 2/1/08, David Ezzio <de...@apache.org> wrote:
> >>> Hi,
> >>>
> >>> As to testing that the EMF objects are useful after serialization,
> >>> this
> >>> is a reasonable expectation of any object that is an instance of the
> >>> Serializable interface.
> >>>
> >>> If it turns out that we don't want to support serializing useful EMF
> >>> objects, then we should take care to remove the Serializable
> >>> interface
> >>> from their implementation.
> >>>
> >>> As I understand it, it is this latter issue that is currently under
> >>> discussion.
> >>>
> >>> An important, but only tangentially related, issue is why the
> >>> build is
> >>> so fragile.
> >>>
> >>> David
> >>>
> >>> Pinaki Poddar wrote:
> >>>> Hi,
> >>>> Serializablity requirement on EMF is arising out of necessity
> >>>> rather
> >>>> than by JPA specifcation.
> >>>> JNDI, too, does not mandate bound objects to be Serializable.
> >>>> But JNDI providers often place that restriction.
> >>>> Non-serializable objects can be bound to JNDI by implementing
> >>>> javax.naming.Referenceable.
> >>>> Kodo's JDO product derivation based on OpenJPA is following this
> >>>> Referenceable route.
> >>>>
> >>>> But I do not see this in openjpa EntityManagerFactory
> >>>> implementation.
> >>>>
> >>>>
> >>>> -----Original Message-----
> >>>> From: Kevin Sutter [mailto:kwsutter@gmail.com]
> >>>> Sent: Friday, February 01, 2008 11:11 AM
> >>>> To: dev@openjpa.apache.org
> >>>> Subject: Re: Failures in TeamCity Derby build and svn commit:
> >>>> r616972
> >>>>
> >>>> On Feb 1, 2008 10:17 AM, Pinaki Poddar <pp...@bea.com> wrote:
> >>>>
> >>>>> Factories needs to be Serializable as they are often placed in
> >>>>> JNDI.
> >>>>
> >>>> The JNDI lookup needs to work for the Factory definition, but it
> >>>> doesn't
> >>>> require that the actual Factory be present in JNDI. And, since
> >>>> the JNDI
> >>>> lookup is meant for the EE (Container) environment, I would guess
> >>>> that
> >>>> most implementations provide a wrapper object in JNDI for the
> >>>> actual
> >>>> Factory.
> >>>>
> >>>> Granted, this wrapper approach is not a requirement. But, I
> >>>> can't find
> >>>> the requirement in the spec that the EMF has to be Serializable
> >>>> either... :-)
> >>>>
> >>>> Kevin
> >>>>
> >>>>
> >>>>> -----Original Message-----
> >>>>> From: Patrick Linskey [mailto:plinskey@gmail.com]
> >>>>> Sent: Thursday, January 31, 2008 7:57 PM
> >>>>> To: dev@openjpa.apache.org
> >>>>> Subject: Re: Failures in TeamCity Derby build and svn commit:
> >>>>> r616972
> >>>>>
> >>>>> To be honest, I don't really understand why a BrokerFactory
> >>>>> needs to
> >>>>> be serializable, so I think probably it'd be good to make sure
> >>>>> that
> >>>>> there is good reason for that before going too deep into this
> >>>>> debugging exercise.
> >>>>>
> >>>>> -Patrick
> >>>>>
> >>>>> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com>
> >>>>> wrote:
> >>>>>>> If the hangup is intolerable, let me know, and I'll back out my
> >>>>>>> change tomorrow.
> >>>>>> Personally, I'd rather have the TeamCity build be clean while you
> >>>>>> investigate. As you noted, the error-reporting from surefire is
> >>>>>> frustratingly inaccurate, and this really erodes the benefit of
> >>>>>> continuous integration.
> >>>>>>
> >>>>>>> It would be handy if I knew a few things. How to pass in VM
> >>>>>>> debug
> >>>>>>> args to Maven for its test runs, where to set the OpenJPA
> >>>>>>> logging
> >>>>>>> when Maven runs the tests, and how to pass in the appropriate
> >>>>>>> parameters when running the test case outside of Maven. Any
> >>>>>>> insights on these questions will be appreciated.
> >>>>>> I usually just add log configuration to the test cases
> >>>>>> themselves,
> >>>>>> in the setUp(...) method:
> >>>>>>
> >>>>>> setUp(Class.class, Class2.class, ..., "openjpa.Log",
> >>>>>> "SQL=TRACE");
> >>>>>>
> >>>>>> I believe that it is also possible to set this at the command
> >>>>>> line
> >>>>>> via
> >>>>>> -Dopenjpa.Log=SQL=TRACE types of clauses.
> >>>>>>
> >>>>>> -Patrick
> >>>>>>
> >>>>>>
> >>>>>> On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
> >>>>>>> Hi,
> >>>>>>>
> >>>>>>> Progress has been slow on the test failures apparently
> >>>>>>> introduced
> >>>>>>> by
> >>>>>>> 616972. I have made progress in reproducing the results.
> >>>>>>>
> >>>>>>> The miscounting in TeamCity is gross. Although it lists 33
> >>>>>>> failing tests complete with a credible stack trace for each, in
> >>>>>>> fact, there is only one failing test.
> >>>>>>>
> >>>>>>> Meanwhile, I believe now, that change 616658 is no longer a
> >>>> suspect.
> >>>>>>> I get very erratic results here, but they seem to vary more by
> >>>>>>> whether I run "mvn test" after "mvn clean" or "mvn install"
> >>>>>>> after
> >>>>>>> "mvn clean". By doing just what TeamCity does, "mvn clean
> >>>>>>> compile
> >>>>>>> test", I can finally reproduce the results locally that TeamCity
> >>>>>>> reports in its Compile-Derby build for this change (395, I
> >>>>>>> think).
> >>>>>>> In spite of this, if I run "mvn
> >>>>>>> -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
> >>>>> test passes.
> >>>>>>> It would be handy if I knew a few things. How to pass in VM
> >>>>>>> debug
> >>>>>>> args to Maven for its test runs, where to set the OpenJPA
> >>>>>>> logging
> >>>>>>> when Maven runs the tests, and how to pass in the appropriate
> >>>>>>> parameters when running the test case outside of Maven. Any
> >>>>>>> insights on these questions will be appreciated.
> >>>>>>>
> >>>>>>> But at least at this point, I can reproduce locally what
> >>>>>>> TeamCity
> >>>>>>> reports, and although it happened only once out of one try, I am
> >>>>>>> optimistic that it is reproducible. I will continue working
> >>>>>>> this
> >>>>>>> tomorrow.
> >>>>>>>
> >>>>>>> If the hangup is intolerable, let me know, and I'll back out my
> >>>>>>> change tomorrow.
> >>>>>>>
> >>>>>>> David Ezzio
> >>>>>>>
> >>>>>>>
> >>>>>>> David Ezzio wrote:
> >>>>>>>> Hi,
> >>>>>>>>
> >>>>>>>> I am convinced that change 616658 (small change to
> >>>>>>>> DBDictionary.java) is the source of the two intermittent
> >>>>>>>> failures
> >>>>> that I am seeing locally.
> >>>>>>>> One of those intermittent failures is the failure in
> >>>>>>>> TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
> >>>>>>>> which is one of the 33 related failures currently seen in
> >>>>>>>> TeamCity
> >>>>> Derby builds.
> >>>>>>>> The other intermittent failure is
> >>>>>>>> TestVersion.testVersionTimestamp
> >>>>>>>> which I reported earlier as occurring locally with change
> >>>> 616658.
> >>>>>>>> When I revert change 616658 locally, both of these issues go
> >>>>>>>> away even with the changes that I submitted to
> >>>>>>>> AbstractBrokerFactory in
> >>>>>>>> change 616972.
> >>>>>>>>
> >>>>>>>> I have a question: is it possible to run a build in TeamCity
> >>>>>>>> without submitting the change (or conditionally submitting the
> >>>>> change) to SVN?
> >>>>>>>> I'd like to find out if reverting 616658 will fix the issue.
> >>>>>>>> If
> >>>>>>>> so, then I can try to figure out why it's a problem.
> >>>>>>>>
> >>>>>>>> Thanks,
> >>>>>>>>
> >>>>>>>> David
> >>>>>>>>
> >>>>>>>> David Ezzio wrote:
> >>>>>>>>> No, but I'm taking a look since my changes appear to be the
> >>>>> cause.
> >>>>>>>>> Patrick Linskey wrote:
> >>>>>>>>>> 5. The tests are now failing in the automated build; any idea
> >>>>>>>>>> why this might be?
> >>>>>>>>>>
> >>>>>>>>>> -Patrick
> >>>>>>>>>>
> >>>>>>>>>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
> >>>>> wrote:
> >>>>>>>>>>> Hi,
> >>>>>>>>>>>
> >>>>>>>>>>> A few questions:
> >>>>>>>>>>>
> >>>>>>>>>>> 1. Were you seeing this cause problems somewhere? I.e., when
> >>>>>>>>>>> do
> >>>>>>>>>>> we serialize BrokerFactory instances?
> >>>>>>>>>>>
> >>>>>>>>>>> 2. Why not just move the initialization code to the
> >>>>>>>>>>> declaration
> >>>>>>>>>>> line (instead of the constructor)?
> >>>>>>>>>>>
> >>>>>>>>>>> 3. It looks like there are other transactional structures
> >>>>>>>>>>> that might not be properly handled in serialization /
> >>>>>>>>>>> deserialization (lifecycle listeners, for example). Should
> >>>>>>>>>>> we
> >>>>> be making more changes there?
> >>>>>>>>>>> -Patrick
> >>>>>>>>>>>
> >>>>>>>>>>>
> >>>>>>>>>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
> >>>>>>>>>>>> Author: dezzio
> >>>>>>>>>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
> >>>>>>>>>>>>
> >>>>>>>>>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
> >>>>>>>>>>>> Log:
> >>>>>>>>>>>> Allow EntityManagerFactory objects to be serialized and
> >>>>>>>>>>>> deserialized successfully.
> >>>>>>>>>>>>
> >>>>>>>>>>>> Added:
> >>>>>>>>>>>>
> >>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
> >>>>>>>>>>>> apa
> >>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>>>>> (with props)
> >>>>>>>>>>>> Modified:
> >>>>>>>>>>>>
> >>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/
> >>>>>>>>>>>> openjp
> >>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
> >>>>>>>>>>>>
> >>>>>>>>>>>>
> >>>>>>>>>>>> Modified:
> >>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/
> >>>>>>>>>>>> openjp
> >>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
> >>>>>>>>>>>>
> >>>>>>>>>>>> URL:
> >>>>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/
> >>>>>>>>>>>> sr
> >>>>>>>>>>>> c/
> >>>>>>>>>>>> main/java/org/apache/openjpa/kernel/
> >>>>>>>>>>>> AbstractBrokerFactory.ja
> >>>>>>>>>>>> va ?rev=616972&r1=616971&r2=616972&view=diff
> >>>>>>>>>>>>
> >>>>>>>>>>>> =
> >>>>>>>>>>>> ===========================================================
> >>>>>>>>>>>> ==
> >>>>>>>>>>>> ================
> >>>>>>>>>>>>
> >>>>>>>>>>>> ---
> >>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/
> >>>>>>>>>>>> openjp
> >>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
> >>>>>>>>>>>> (original)
> >>>>>>>>>>>> +++
> >>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/
> >>>>>>>>>>>> openjp
> >>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java Wed Jan 30 16:59:02
> >>>>>>>>>>>> 2008 @@ -147,8 +147,7 @@
> >>>>>>>>>>>> */
> >>>>>>>>>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
> >>>>> config) {
> >>>>>>>>>>>> _conf = config;
> >>>>>>>>>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
> >>>>>>>>>>>> - ConcurrentReferenceHashSet.WEAK);
> >>>>>>>>>>>> + getPcClassLoaders();
> >>>>>>>>>>>> }
> >>>>>>>>>>>>
> >>>>>>>>>>>> /**
> >>>>>>>>>>>> @@ -287,13 +286,13 @@
> >>>>>>>>>>>> if (needsSub(cls))
> >>>>>>>>>>>> toRedefine.add(cls);
> >>>>>>>>>>>> }
> >>>>>>>>>>>> - _pcClassLoaders.add(loader);
> >>>>>>>>>>>> + getPcClassLoaders().add(loader);
> >>>>>>>>>>>> _pcClassNames = c;
> >>>>>>>>>>>> }
> >>>>>>>>>>>> _persistentTypesLoaded = true;
> >>>>>>>>>>>> } else {
> >>>>>>>>>>>> // reload with this loader
> >>>>>>>>>>>> - if (_pcClassLoaders.add(loader)) {
> >>>>>>>>>>>> + if (getPcClassLoaders().add(loader)) {
> >>>>>>>>>>>> for (Iterator itr =
> >>>>>>>>>>>> _pcClassNames.iterator();
> >>>>>>>>>>>> itr.hasNext();) {
> >>>>>>>>>>>> try {
> >>>>>>>>>>>> Class cls = @@ -818,4 +817,15 @@
> >>>>>>>>>>>> _transactional.remove (_trans);
> >>>>>>>>>>>> }
> >>>>>>>>>>>> }
> >>>>>>>>>>>> +
> >>>>>>>>>>>> + /**
> >>>>>>>>>>>> + * Method insures that deserialized EMF has this
> >>>>>>>>>>>> + reference
> >>>>>>>>>>>> re-instantiated
> >>>>>>>>>>>> + */
> >>>>>>>>>>>> + private Collection getPcClassLoaders() {
> >>>>>>>>>>>> + if (_pcClassLoaders == null)
> >>>>>>>>>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
> >>>>>>>>>>>> + ConcurrentReferenceHashSet.WEAK);
> >>>>>>>>>>>> +
> >>>>>>>>>>>> + return _pcClassLoaders;
> >>>>>>>>>>>> + }
> >>>>>>>>>>>> }
> >>>>>>>>>>>>
> >>>>>>>>>>>> Added:
> >>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
> >>>>>>>>>>>> apa
> >>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>>>>>
> >>>>>>>>>>>> URL:
> >>>>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-
> >>>>>>>>>>>> persisten
> >>>>>>>>>>>> ce
> >>>>>>>>>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/
> >>>>>>>>>>>> Te
> >>>>>>>>>>>> st SerializedFactory.java?rev=616972&view=auto
> >>>>>>>>>>>>
> >>>>>>>>>>>> =
> >>>>>>>>>>>> ===========================================================
> >>>>>>>>>>>> ==
> >>>>>>>>>>>> ================
> >>>>>>>>>>>>
> >>>>>>>>>>>> ---
> >>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
> >>>>>>>>>>>> apa
> >>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>>>>> (added)
> >>>>>>>>>>>> +++
> >>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
> >>>>>>>>>>>> apa
> >>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>>>>> Wed Jan 30 16:59:02 2008
> >>>>>>>>>>>> @@ -0,0 +1,74 @@
> >>>>>>>>>>>> +/*
> >>>>>>>>>>>> + * Licensed to the Apache Software Foundation (ASF) under
> >>>>>>>>>>>> +one
> >>>>>>>>>>>> + * or more contributor license agreements. See the NOTICE
> >>>>>>>>>>>> +file
> >>>>>>>>>>>> + * distributed with this work for additional information
> >>>>>>>>>>>> + * regarding copyright ownership. The ASF licenses this
> >>>>>>>>>>>> +file
> >>>>>>>>>>>> + * to you 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 org.apache.openjpa.persistence.simple;
> >>>>>>>>>>>> +
> >>>>>>>>>>>> +import java.io.*;
> >>>>>>>>>>>> +
> >>>>>>>>>>>> +import javax.persistence.EntityManager; import
> >>>>>>>>>>>> +javax.persistence.EntityManagerFactory;
> >>>>>>>>>>>> +import javax.persistence.EntityTransaction;
> >>>>>>>>>>>> +
> >>>>>>>>>>>> +import junit.textui.TestRunner; import
> >>>>>>>>>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
> >>>>>>>>>>>> +import
> >>>>>>>>>>>> +org.apache.openjpa.persistence.test.SingleEMFTestCase;
> >>>>>>>>>>>> +
> >>>>>>>>>>>> +/**
> >>>>>>>>>>>> + * Tests that a EntityManagerFactory can be used after
> >>>>> serialization.
> >>>>>>>>>>>> + *
> >>>>>>>>>>>> + * @author David Ezzio
> >>>>>>>>>>>> + */
> >>>>>>>>>>>> +public class TestSerializedFactory
> >>>>>>>>>>>> + extends SingleEMFTestCase {
> >>>>>>>>>>>> +
> >>>>>>>>>>>> + public void setUp() {
> >>>>>>>>>>>> + setUp(AllFieldTypes.class);
> >>>>>>>>>>>> + }
> >>>>>>>>>>>> +
> >>>>>>>>>>>> + public void testSerializedEntityManagerFactory()
> >>>>>>>>>>>> throws
> >>>>>>>>>>>> Exception {
> >>>>>>>>>>>> + // serialize and deserialize the entity manager
> >>>>> factory
> >>>>>>>>>>>> + ByteArrayOutputStream baos = new
> >>>>> ByteArrayOutputStream();
> >>>>>>>>>>>> + ObjectOutputStream oos = new
> >>>>> ObjectOutputStream(baos);
> >>>>>>>>>>>> + oos.writeObject(emf);
> >>>>>>>>>>>> + EntityManagerFactory emf2 =
> >>>>>>>>>>>> + (EntityManagerFactory) new ObjectInputStream(
> >>>>>>>>>>>> + new
> >>>>>>>>>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
> >>>>>>>>>>>> +
> >>>>>>>>>>>> + // use the deserialized entity manager factory
> >>>>>>>>>>>> + assertTrue("The deserialized entity manager
> >>>>>>>>>>>> factory
> >>>>>>>>>>>> + is not
> >>>>>>>>>>>> open",
> >>>>>>>>>>>> + emf2.isOpen());
> >>>>>>>>>>>> + EntityManager em = emf2.createEntityManager();
> >>>>>>>>>>>> + assertTrue("The newly created entity manager is
> >>>>>>>>>>>> not
> >>>>>>>>>>>> + open",
> >>>>>>>>>>>> em.isOpen());
> >>>>>>>>>>>> +
> >>>>>>>>>>>> + // exercise the entity manager produced from the
> >>>>>>>>>>>> deserialized EMF
> >>>>>>>>>>>> + em.getTransaction().begin();
> >>>>>>>>>>>> + em.persist(new AllFieldTypes());
> >>>>>>>>>>>> + em.getTransaction().commit();
> >>>>>>>>>>>> +
> >>>>>>>>>>>> + // close the extra resources
> >>>>>>>>>>>> + em.close();
> >>>>>>>>>>>> + assertFalse("The entity manager is not closed",
> >>>>> em.isOpen());
> >>>>>>>>>>>> + emf2.close();
> >>>>>>>>>>>> + assertFalse("The entity manager factory is not
> >>>>>>>>>>>> + closed",
> >>>>>>>>>>>> emf2.isOpen());
> >>>>>>>>>>>> + }
> >>>>>>>>>>>> +
> >>>>>>>>>>>> + public static void main(String[] args) {
> >>>>>>>>>>>> + TestRunner.run(TestSerializedFactory.class);
> >>>>>>>>>>>> + }
> >>>>>>>>>>>> +}
> >>>>>>>>>>>> +
> >>>>>>>>>>>>
> >>>>>>>>>>>> Propchange:
> >>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
> >>>>>>>>>>>> apa
> >>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>>>>>
> >>>>>>>>>>>> ------------------------------------------------------------
> >>>>>>>>>>>> --
> >>>>>>>>>>>> ----------------
> >>>>>>>>>>>>
> >>>>>>>>>>>> svn:eol-style = native
> >>>>>>>>>>>>
> >>>>>>>>>>>>
> >>>>>>>>>>>>
> >>>>>>>>>>> --
> >>>>>>>>>>> Patrick Linskey
> >>>>>>>>>>> 202 669 5907
> >>>>>>>>>>>
> >>>>>>>>>>
> >>>>>>
> >>>>>> --
> >>>>>> Patrick Linskey
> >>>>>> 202 669 5907
> >>>>>>
> >>>>>
> >>>>> --
> >>>>> Patrick Linskey
> >>>>> 202 669 5907
> >>>>>
> >>>>> Notice: This email message, together with any attachments, may
> >>>>> contain information of BEA Systems, Inc., its subsidiaries
> >>>>> and
> >>>>> affiliated entities, that may be confidential, proprietary,
> >>>>> copyrighted and/or legally privileged, and is intended solely
> >>>>> for the
> >>>>> use of the individual or entity named in this message. If you
> >>>>> are not
> >>>>> the intended recipient, and have received this message in error,
> >>>>> please immediately return this by email and then delete it.
> >>>>>
> >>>> Notice: This email message, together with any attachments, may
> >>>> contain
> >>> information of BEA Systems, Inc., its subsidiaries and
> >>> affiliated
> >>> entities, that may be confidential, proprietary, copyrighted
> >>> and/or
> >>> legally privileged, and is intended solely for the use of the
> >>> individual or
> >>> entity named in this message. If you are not the intended
> >>> recipient, and
> >>> have received this message in error, please immediately return
> >>> this by email
> >>> and then delete it.
>
> Craig Russell
> Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
> 408 276-5638 mailto:Craig.Russell@sun.com
> P.S. A good JDO? O, Gasp!
>
>
--
Patrick Linskey
202 669 5907
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by Craig L Russell <Cr...@Sun.COM>.
I think we should ask the JPA EG to resolve whether EMF should be
serializable.
When a session bean is passivated of saved for failover, it is
serialized. When serializing the state of a session bean, the fields
of the bean need to be serialized. If the EMF is stored in a field,
then it needs to be restored when the session is restored. It would be
nice if the semantics of passivation and save for failover of an EMF
were standardized.
Craig
On Feb 5, 2008, at 5:01 PM, David Ezzio wrote:
> Hi Patrick,
>
> The build was fragile because a couple of test cases were fragile
> (one did not clean out data before checking whether only the data it
> added was present, and the other did not ensure that time passed
> between updates when using a timestamp version check), and because
> the testing framework itself was fragile in that it attempted to
> delete data using a new entity manager when transactions might
> remain open (due to test failures), resulting in deadlock in the
> database. I've fixed these issues.
>
> Since there is no present resolution of the question whether to
> serialize or not the EMF, I have put in my change once again
> together with the updated test case. I expect that TeamCity will
> now be happy due to the aforementioned fixes of the testing
> framework, etc.
>
> Of course, we can always revisit the issue of whether to allow the
> EMF to be serialized, but since it currently is serializable, I put
> in the fix to actually allow it to serialize successfully.
>
> David
>
> Patrick Linskey wrote:
>> Yes, I agree that if we make serialization not work, we should make
>> the BF not implement Serializable. That is my preferred approach:
>> even
>> if Kodo needs to serialize a PMF for JNDI purposes or something, we
>> could do that in the PMF impl, instead of in the BF.
>> IMO, the build is fragile because it runs multiple unrelated tests in
>> the same VM with no strong isolation between tests. This leads to
>> unexpected side effects sometimes. This can be frustrating, but the
>> benefits are considerable: faster test run time, plus we get to find
>> out about weird unintended side effects.
>> -Patrick
>> On 2/1/08, David Ezzio <de...@apache.org> wrote:
>>> Hi,
>>>
>>> As to testing that the EMF objects are useful after serialization,
>>> this
>>> is a reasonable expectation of any object that is an instance of the
>>> Serializable interface.
>>>
>>> If it turns out that we don't want to support serializing useful EMF
>>> objects, then we should take care to remove the Serializable
>>> interface
>>> from their implementation.
>>>
>>> As I understand it, it is this latter issue that is currently under
>>> discussion.
>>>
>>> An important, but only tangentially related, issue is why the
>>> build is
>>> so fragile.
>>>
>>> David
>>>
>>> Pinaki Poddar wrote:
>>>> Hi,
>>>> Serializablity requirement on EMF is arising out of necessity
>>>> rather
>>>> than by JPA specifcation.
>>>> JNDI, too, does not mandate bound objects to be Serializable.
>>>> But JNDI providers often place that restriction.
>>>> Non-serializable objects can be bound to JNDI by implementing
>>>> javax.naming.Referenceable.
>>>> Kodo's JDO product derivation based on OpenJPA is following this
>>>> Referenceable route.
>>>>
>>>> But I do not see this in openjpa EntityManagerFactory
>>>> implementation.
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: Kevin Sutter [mailto:kwsutter@gmail.com]
>>>> Sent: Friday, February 01, 2008 11:11 AM
>>>> To: dev@openjpa.apache.org
>>>> Subject: Re: Failures in TeamCity Derby build and svn commit:
>>>> r616972
>>>>
>>>> On Feb 1, 2008 10:17 AM, Pinaki Poddar <pp...@bea.com> wrote:
>>>>
>>>>> Factories needs to be Serializable as they are often placed in
>>>>> JNDI.
>>>>
>>>> The JNDI lookup needs to work for the Factory definition, but it
>>>> doesn't
>>>> require that the actual Factory be present in JNDI. And, since
>>>> the JNDI
>>>> lookup is meant for the EE (Container) environment, I would guess
>>>> that
>>>> most implementations provide a wrapper object in JNDI for the
>>>> actual
>>>> Factory.
>>>>
>>>> Granted, this wrapper approach is not a requirement. But, I
>>>> can't find
>>>> the requirement in the spec that the EMF has to be Serializable
>>>> either... :-)
>>>>
>>>> Kevin
>>>>
>>>>
>>>>> -----Original Message-----
>>>>> From: Patrick Linskey [mailto:plinskey@gmail.com]
>>>>> Sent: Thursday, January 31, 2008 7:57 PM
>>>>> To: dev@openjpa.apache.org
>>>>> Subject: Re: Failures in TeamCity Derby build and svn commit:
>>>>> r616972
>>>>>
>>>>> To be honest, I don't really understand why a BrokerFactory
>>>>> needs to
>>>>> be serializable, so I think probably it'd be good to make sure
>>>>> that
>>>>> there is good reason for that before going too deep into this
>>>>> debugging exercise.
>>>>>
>>>>> -Patrick
>>>>>
>>>>> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com>
>>>>> wrote:
>>>>>>> If the hangup is intolerable, let me know, and I'll back out my
>>>>>>> change tomorrow.
>>>>>> Personally, I'd rather have the TeamCity build be clean while you
>>>>>> investigate. As you noted, the error-reporting from surefire is
>>>>>> frustratingly inaccurate, and this really erodes the benefit of
>>>>>> continuous integration.
>>>>>>
>>>>>>> It would be handy if I knew a few things. How to pass in VM
>>>>>>> debug
>>>>>>> args to Maven for its test runs, where to set the OpenJPA
>>>>>>> logging
>>>>>>> when Maven runs the tests, and how to pass in the appropriate
>>>>>>> parameters when running the test case outside of Maven. Any
>>>>>>> insights on these questions will be appreciated.
>>>>>> I usually just add log configuration to the test cases
>>>>>> themselves,
>>>>>> in the setUp(...) method:
>>>>>>
>>>>>> setUp(Class.class, Class2.class, ..., "openjpa.Log",
>>>>>> "SQL=TRACE");
>>>>>>
>>>>>> I believe that it is also possible to set this at the command
>>>>>> line
>>>>>> via
>>>>>> -Dopenjpa.Log=SQL=TRACE types of clauses.
>>>>>>
>>>>>> -Patrick
>>>>>>
>>>>>>
>>>>>> On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> Progress has been slow on the test failures apparently
>>>>>>> introduced
>>>>>>> by
>>>>>>> 616972. I have made progress in reproducing the results.
>>>>>>>
>>>>>>> The miscounting in TeamCity is gross. Although it lists 33
>>>>>>> failing tests complete with a credible stack trace for each, in
>>>>>>> fact, there is only one failing test.
>>>>>>>
>>>>>>> Meanwhile, I believe now, that change 616658 is no longer a
>>>> suspect.
>>>>>>> I get very erratic results here, but they seem to vary more by
>>>>>>> whether I run "mvn test" after "mvn clean" or "mvn install"
>>>>>>> after
>>>>>>> "mvn clean". By doing just what TeamCity does, "mvn clean
>>>>>>> compile
>>>>>>> test", I can finally reproduce the results locally that TeamCity
>>>>>>> reports in its Compile-Derby build for this change (395, I
>>>>>>> think).
>>>>>>> In spite of this, if I run "mvn
>>>>>>> -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
>>>>> test passes.
>>>>>>> It would be handy if I knew a few things. How to pass in VM
>>>>>>> debug
>>>>>>> args to Maven for its test runs, where to set the OpenJPA
>>>>>>> logging
>>>>>>> when Maven runs the tests, and how to pass in the appropriate
>>>>>>> parameters when running the test case outside of Maven. Any
>>>>>>> insights on these questions will be appreciated.
>>>>>>>
>>>>>>> But at least at this point, I can reproduce locally what
>>>>>>> TeamCity
>>>>>>> reports, and although it happened only once out of one try, I am
>>>>>>> optimistic that it is reproducible. I will continue working
>>>>>>> this
>>>>>>> tomorrow.
>>>>>>>
>>>>>>> If the hangup is intolerable, let me know, and I'll back out my
>>>>>>> change tomorrow.
>>>>>>>
>>>>>>> David Ezzio
>>>>>>>
>>>>>>>
>>>>>>> David Ezzio wrote:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> I am convinced that change 616658 (small change to
>>>>>>>> DBDictionary.java) is the source of the two intermittent
>>>>>>>> failures
>>>>> that I am seeing locally.
>>>>>>>> One of those intermittent failures is the failure in
>>>>>>>> TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
>>>>>>>> which is one of the 33 related failures currently seen in
>>>>>>>> TeamCity
>>>>> Derby builds.
>>>>>>>> The other intermittent failure is
>>>>>>>> TestVersion.testVersionTimestamp
>>>>>>>> which I reported earlier as occurring locally with change
>>>> 616658.
>>>>>>>> When I revert change 616658 locally, both of these issues go
>>>>>>>> away even with the changes that I submitted to
>>>>>>>> AbstractBrokerFactory in
>>>>>>>> change 616972.
>>>>>>>>
>>>>>>>> I have a question: is it possible to run a build in TeamCity
>>>>>>>> without submitting the change (or conditionally submitting the
>>>>> change) to SVN?
>>>>>>>> I'd like to find out if reverting 616658 will fix the issue.
>>>>>>>> If
>>>>>>>> so, then I can try to figure out why it's a problem.
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>>
>>>>>>>> David
>>>>>>>>
>>>>>>>> David Ezzio wrote:
>>>>>>>>> No, but I'm taking a look since my changes appear to be the
>>>>> cause.
>>>>>>>>> Patrick Linskey wrote:
>>>>>>>>>> 5. The tests are now failing in the automated build; any idea
>>>>>>>>>> why this might be?
>>>>>>>>>>
>>>>>>>>>> -Patrick
>>>>>>>>>>
>>>>>>>>>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
>>>>> wrote:
>>>>>>>>>>> Hi,
>>>>>>>>>>>
>>>>>>>>>>> A few questions:
>>>>>>>>>>>
>>>>>>>>>>> 1. Were you seeing this cause problems somewhere? I.e., when
>>>>>>>>>>> do
>>>>>>>>>>> we serialize BrokerFactory instances?
>>>>>>>>>>>
>>>>>>>>>>> 2. Why not just move the initialization code to the
>>>>>>>>>>> declaration
>>>>>>>>>>> line (instead of the constructor)?
>>>>>>>>>>>
>>>>>>>>>>> 3. It looks like there are other transactional structures
>>>>>>>>>>> that might not be properly handled in serialization /
>>>>>>>>>>> deserialization (lifecycle listeners, for example). Should
>>>>>>>>>>> we
>>>>> be making more changes there?
>>>>>>>>>>> -Patrick
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
>>>>>>>>>>>> Author: dezzio
>>>>>>>>>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
>>>>>>>>>>>>
>>>>>>>>>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
>>>>>>>>>>>> Log:
>>>>>>>>>>>> Allow EntityManagerFactory objects to be serialized and
>>>>>>>>>>>> deserialized successfully.
>>>>>>>>>>>>
>>>>>>>>>>>> Added:
>>>>>>>>>>>>
>>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
>>>>>>>>>>>> apa
>>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>>> (with props)
>>>>>>>>>>>> Modified:
>>>>>>>>>>>>
>>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/
>>>>>>>>>>>> openjp
>>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> Modified:
>>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/
>>>>>>>>>>>> openjp
>>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>>>>
>>>>>>>>>>>> URL:
>>>>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/
>>>>>>>>>>>> sr
>>>>>>>>>>>> c/
>>>>>>>>>>>> main/java/org/apache/openjpa/kernel/
>>>>>>>>>>>> AbstractBrokerFactory.ja
>>>>>>>>>>>> va ?rev=616972&r1=616971&r2=616972&view=diff
>>>>>>>>>>>>
>>>>>>>>>>>> =
>>>>>>>>>>>> ===========================================================
>>>>>>>>>>>> ==
>>>>>>>>>>>> ================
>>>>>>>>>>>>
>>>>>>>>>>>> ---
>>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/
>>>>>>>>>>>> openjp
>>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>>>> (original)
>>>>>>>>>>>> +++
>>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/
>>>>>>>>>>>> openjp
>>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java Wed Jan 30 16:59:02
>>>>>>>>>>>> 2008 @@ -147,8 +147,7 @@
>>>>>>>>>>>> */
>>>>>>>>>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
>>>>> config) {
>>>>>>>>>>>> _conf = config;
>>>>>>>>>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>>>>>>>>> - ConcurrentReferenceHashSet.WEAK);
>>>>>>>>>>>> + getPcClassLoaders();
>>>>>>>>>>>> }
>>>>>>>>>>>>
>>>>>>>>>>>> /**
>>>>>>>>>>>> @@ -287,13 +286,13 @@
>>>>>>>>>>>> if (needsSub(cls))
>>>>>>>>>>>> toRedefine.add(cls);
>>>>>>>>>>>> }
>>>>>>>>>>>> - _pcClassLoaders.add(loader);
>>>>>>>>>>>> + getPcClassLoaders().add(loader);
>>>>>>>>>>>> _pcClassNames = c;
>>>>>>>>>>>> }
>>>>>>>>>>>> _persistentTypesLoaded = true;
>>>>>>>>>>>> } else {
>>>>>>>>>>>> // reload with this loader
>>>>>>>>>>>> - if (_pcClassLoaders.add(loader)) {
>>>>>>>>>>>> + if (getPcClassLoaders().add(loader)) {
>>>>>>>>>>>> for (Iterator itr =
>>>>>>>>>>>> _pcClassNames.iterator();
>>>>>>>>>>>> itr.hasNext();) {
>>>>>>>>>>>> try {
>>>>>>>>>>>> Class cls = @@ -818,4 +817,15 @@
>>>>>>>>>>>> _transactional.remove (_trans);
>>>>>>>>>>>> }
>>>>>>>>>>>> }
>>>>>>>>>>>> +
>>>>>>>>>>>> + /**
>>>>>>>>>>>> + * Method insures that deserialized EMF has this
>>>>>>>>>>>> + reference
>>>>>>>>>>>> re-instantiated
>>>>>>>>>>>> + */
>>>>>>>>>>>> + private Collection getPcClassLoaders() {
>>>>>>>>>>>> + if (_pcClassLoaders == null)
>>>>>>>>>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>>>>>>>>> + ConcurrentReferenceHashSet.WEAK);
>>>>>>>>>>>> +
>>>>>>>>>>>> + return _pcClassLoaders;
>>>>>>>>>>>> + }
>>>>>>>>>>>> }
>>>>>>>>>>>>
>>>>>>>>>>>> Added:
>>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
>>>>>>>>>>>> apa
>>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>>>
>>>>>>>>>>>> URL:
>>>>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-
>>>>>>>>>>>> persisten
>>>>>>>>>>>> ce
>>>>>>>>>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/
>>>>>>>>>>>> Te
>>>>>>>>>>>> st SerializedFactory.java?rev=616972&view=auto
>>>>>>>>>>>>
>>>>>>>>>>>> =
>>>>>>>>>>>> ===========================================================
>>>>>>>>>>>> ==
>>>>>>>>>>>> ================
>>>>>>>>>>>>
>>>>>>>>>>>> ---
>>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
>>>>>>>>>>>> apa
>>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>>> (added)
>>>>>>>>>>>> +++
>>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
>>>>>>>>>>>> apa
>>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>>> Wed Jan 30 16:59:02 2008
>>>>>>>>>>>> @@ -0,0 +1,74 @@
>>>>>>>>>>>> +/*
>>>>>>>>>>>> + * Licensed to the Apache Software Foundation (ASF) under
>>>>>>>>>>>> +one
>>>>>>>>>>>> + * or more contributor license agreements. See the NOTICE
>>>>>>>>>>>> +file
>>>>>>>>>>>> + * distributed with this work for additional information
>>>>>>>>>>>> + * regarding copyright ownership. The ASF licenses this
>>>>>>>>>>>> +file
>>>>>>>>>>>> + * to you 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 org.apache.openjpa.persistence.simple;
>>>>>>>>>>>> +
>>>>>>>>>>>> +import java.io.*;
>>>>>>>>>>>> +
>>>>>>>>>>>> +import javax.persistence.EntityManager; import
>>>>>>>>>>>> +javax.persistence.EntityManagerFactory;
>>>>>>>>>>>> +import javax.persistence.EntityTransaction;
>>>>>>>>>>>> +
>>>>>>>>>>>> +import junit.textui.TestRunner; import
>>>>>>>>>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
>>>>>>>>>>>> +import
>>>>>>>>>>>> +org.apache.openjpa.persistence.test.SingleEMFTestCase;
>>>>>>>>>>>> +
>>>>>>>>>>>> +/**
>>>>>>>>>>>> + * Tests that a EntityManagerFactory can be used after
>>>>> serialization.
>>>>>>>>>>>> + *
>>>>>>>>>>>> + * @author David Ezzio
>>>>>>>>>>>> + */
>>>>>>>>>>>> +public class TestSerializedFactory
>>>>>>>>>>>> + extends SingleEMFTestCase {
>>>>>>>>>>>> +
>>>>>>>>>>>> + public void setUp() {
>>>>>>>>>>>> + setUp(AllFieldTypes.class);
>>>>>>>>>>>> + }
>>>>>>>>>>>> +
>>>>>>>>>>>> + public void testSerializedEntityManagerFactory()
>>>>>>>>>>>> throws
>>>>>>>>>>>> Exception {
>>>>>>>>>>>> + // serialize and deserialize the entity manager
>>>>> factory
>>>>>>>>>>>> + ByteArrayOutputStream baos = new
>>>>> ByteArrayOutputStream();
>>>>>>>>>>>> + ObjectOutputStream oos = new
>>>>> ObjectOutputStream(baos);
>>>>>>>>>>>> + oos.writeObject(emf);
>>>>>>>>>>>> + EntityManagerFactory emf2 =
>>>>>>>>>>>> + (EntityManagerFactory) new ObjectInputStream(
>>>>>>>>>>>> + new
>>>>>>>>>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
>>>>>>>>>>>> +
>>>>>>>>>>>> + // use the deserialized entity manager factory
>>>>>>>>>>>> + assertTrue("The deserialized entity manager
>>>>>>>>>>>> factory
>>>>>>>>>>>> + is not
>>>>>>>>>>>> open",
>>>>>>>>>>>> + emf2.isOpen());
>>>>>>>>>>>> + EntityManager em = emf2.createEntityManager();
>>>>>>>>>>>> + assertTrue("The newly created entity manager is
>>>>>>>>>>>> not
>>>>>>>>>>>> + open",
>>>>>>>>>>>> em.isOpen());
>>>>>>>>>>>> +
>>>>>>>>>>>> + // exercise the entity manager produced from the
>>>>>>>>>>>> deserialized EMF
>>>>>>>>>>>> + em.getTransaction().begin();
>>>>>>>>>>>> + em.persist(new AllFieldTypes());
>>>>>>>>>>>> + em.getTransaction().commit();
>>>>>>>>>>>> +
>>>>>>>>>>>> + // close the extra resources
>>>>>>>>>>>> + em.close();
>>>>>>>>>>>> + assertFalse("The entity manager is not closed",
>>>>> em.isOpen());
>>>>>>>>>>>> + emf2.close();
>>>>>>>>>>>> + assertFalse("The entity manager factory is not
>>>>>>>>>>>> + closed",
>>>>>>>>>>>> emf2.isOpen());
>>>>>>>>>>>> + }
>>>>>>>>>>>> +
>>>>>>>>>>>> + public static void main(String[] args) {
>>>>>>>>>>>> + TestRunner.run(TestSerializedFactory.class);
>>>>>>>>>>>> + }
>>>>>>>>>>>> +}
>>>>>>>>>>>> +
>>>>>>>>>>>>
>>>>>>>>>>>> Propchange:
>>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/
>>>>>>>>>>>> apa
>>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>>>
>>>>>>>>>>>> ------------------------------------------------------------
>>>>>>>>>>>> --
>>>>>>>>>>>> ----------------
>>>>>>>>>>>>
>>>>>>>>>>>> svn:eol-style = native
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>> --
>>>>>>>>>>> Patrick Linskey
>>>>>>>>>>> 202 669 5907
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>
>>>>>> --
>>>>>> Patrick Linskey
>>>>>> 202 669 5907
>>>>>>
>>>>>
>>>>> --
>>>>> Patrick Linskey
>>>>> 202 669 5907
>>>>>
>>>>> Notice: This email message, together with any attachments, may
>>>>> contain information of BEA Systems, Inc., its subsidiaries
>>>>> and
>>>>> affiliated entities, that may be confidential, proprietary,
>>>>> copyrighted and/or legally privileged, and is intended solely
>>>>> for the
>>>>> use of the individual or entity named in this message. If you
>>>>> are not
>>>>> the intended recipient, and have received this message in error,
>>>>> please immediately return this by email and then delete it.
>>>>>
>>>> Notice: This email message, together with any attachments, may
>>>> contain
>>> information of BEA Systems, Inc., its subsidiaries and
>>> affiliated
>>> entities, that may be confidential, proprietary, copyrighted
>>> and/or
>>> legally privileged, and is intended solely for the use of the
>>> individual or
>>> entity named in this message. If you are not the intended
>>> recipient, and
>>> have received this message in error, please immediately return
>>> this by email
>>> and then delete it.
Craig Russell
Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
408 276-5638 mailto:Craig.Russell@sun.com
P.S. A good JDO? O, Gasp!
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by David Ezzio <de...@apache.org>.
Hi Patrick,
The build was fragile because a couple of test cases were fragile (one
did not clean out data before checking whether only the data it added
was present, and the other did not ensure that time passed between
updates when using a timestamp version check), and because the testing
framework itself was fragile in that it attempted to delete data using a
new entity manager when transactions might remain open (due to test
failures), resulting in deadlock in the database. I've fixed these
issues.
Since there is no present resolution of the question whether to
serialize or not the EMF, I have put in my change once again together
with the updated test case. I expect that TeamCity will now be happy
due to the aforementioned fixes of the testing framework, etc.
Of course, we can always revisit the issue of whether to allow the EMF
to be serialized, but since it currently is serializable, I put in the
fix to actually allow it to serialize successfully.
David
Patrick Linskey wrote:
> Yes, I agree that if we make serialization not work, we should make
> the BF not implement Serializable. That is my preferred approach: even
> if Kodo needs to serialize a PMF for JNDI purposes or something, we
> could do that in the PMF impl, instead of in the BF.
>
> IMO, the build is fragile because it runs multiple unrelated tests in
> the same VM with no strong isolation between tests. This leads to
> unexpected side effects sometimes. This can be frustrating, but the
> benefits are considerable: faster test run time, plus we get to find
> out about weird unintended side effects.
>
> -Patrick
>
> On 2/1/08, David Ezzio <de...@apache.org> wrote:
>> Hi,
>>
>> As to testing that the EMF objects are useful after serialization, this
>> is a reasonable expectation of any object that is an instance of the
>> Serializable interface.
>>
>> If it turns out that we don't want to support serializing useful EMF
>> objects, then we should take care to remove the Serializable interface
>> from their implementation.
>>
>> As I understand it, it is this latter issue that is currently under
>> discussion.
>>
>> An important, but only tangentially related, issue is why the build is
>> so fragile.
>>
>> David
>>
>> Pinaki Poddar wrote:
>>> Hi,
>>> Serializablity requirement on EMF is arising out of necessity rather
>>> than by JPA specifcation.
>>> JNDI, too, does not mandate bound objects to be Serializable.
>>> But JNDI providers often place that restriction.
>>> Non-serializable objects can be bound to JNDI by implementing
>>> javax.naming.Referenceable.
>>> Kodo's JDO product derivation based on OpenJPA is following this
>>> Referenceable route.
>>>
>>> But I do not see this in openjpa EntityManagerFactory implementation.
>>>
>>>
>>> -----Original Message-----
>>> From: Kevin Sutter [mailto:kwsutter@gmail.com]
>>> Sent: Friday, February 01, 2008 11:11 AM
>>> To: dev@openjpa.apache.org
>>> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
>>>
>>> On Feb 1, 2008 10:17 AM, Pinaki Poddar <pp...@bea.com> wrote:
>>>
>>>> Factories needs to be Serializable as they are often placed in JNDI.
>>>
>>> The JNDI lookup needs to work for the Factory definition, but it doesn't
>>> require that the actual Factory be present in JNDI. And, since the JNDI
>>> lookup is meant for the EE (Container) environment, I would guess that
>>> most implementations provide a wrapper object in JNDI for the actual
>>> Factory.
>>>
>>> Granted, this wrapper approach is not a requirement. But, I can't find
>>> the requirement in the spec that the EMF has to be Serializable
>>> either... :-)
>>>
>>> Kevin
>>>
>>>
>>>> -----Original Message-----
>>>> From: Patrick Linskey [mailto:plinskey@gmail.com]
>>>> Sent: Thursday, January 31, 2008 7:57 PM
>>>> To: dev@openjpa.apache.org
>>>> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
>>>>
>>>> To be honest, I don't really understand why a BrokerFactory needs to
>>>> be serializable, so I think probably it'd be good to make sure that
>>>> there is good reason for that before going too deep into this
>>>> debugging exercise.
>>>>
>>>> -Patrick
>>>>
>>>> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
>>>>>> If the hangup is intolerable, let me know, and I'll back out my
>>>>>> change tomorrow.
>>>>> Personally, I'd rather have the TeamCity build be clean while you
>>>>> investigate. As you noted, the error-reporting from surefire is
>>>>> frustratingly inaccurate, and this really erodes the benefit of
>>>>> continuous integration.
>>>>>
>>>>>> It would be handy if I knew a few things. How to pass in VM debug
>>>>>> args to Maven for its test runs, where to set the OpenJPA logging
>>>>>> when Maven runs the tests, and how to pass in the appropriate
>>>>>> parameters when running the test case outside of Maven. Any
>>>>>> insights on these questions will be appreciated.
>>>>> I usually just add log configuration to the test cases themselves,
>>>>> in the setUp(...) method:
>>>>>
>>>>> setUp(Class.class, Class2.class, ..., "openjpa.Log",
>>>>> "SQL=TRACE");
>>>>>
>>>>> I believe that it is also possible to set this at the command line
>>>>> via
>>>>> -Dopenjpa.Log=SQL=TRACE types of clauses.
>>>>>
>>>>> -Patrick
>>>>>
>>>>>
>>>>> On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Progress has been slow on the test failures apparently introduced
>>>>>> by
>>>>>> 616972. I have made progress in reproducing the results.
>>>>>>
>>>>>> The miscounting in TeamCity is gross. Although it lists 33
>>>>>> failing tests complete with a credible stack trace for each, in
>>>>>> fact, there is only one failing test.
>>>>>>
>>>>>> Meanwhile, I believe now, that change 616658 is no longer a
>>> suspect.
>>>>>> I get very erratic results here, but they seem to vary more by
>>>>>> whether I run "mvn test" after "mvn clean" or "mvn install" after
>>>>>> "mvn clean". By doing just what TeamCity does, "mvn clean compile
>>>>>> test", I can finally reproduce the results locally that TeamCity
>>>>>> reports in its Compile-Derby build for this change (395, I think).
>>>>>> In spite of this, if I run "mvn
>>>>>> -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
>>>> test passes.
>>>>>> It would be handy if I knew a few things. How to pass in VM debug
>>>>>> args to Maven for its test runs, where to set the OpenJPA logging
>>>>>> when Maven runs the tests, and how to pass in the appropriate
>>>>>> parameters when running the test case outside of Maven. Any
>>>>>> insights on these questions will be appreciated.
>>>>>>
>>>>>> But at least at this point, I can reproduce locally what TeamCity
>>>>>> reports, and although it happened only once out of one try, I am
>>>>>> optimistic that it is reproducible. I will continue working this
>>>>>> tomorrow.
>>>>>>
>>>>>> If the hangup is intolerable, let me know, and I'll back out my
>>>>>> change tomorrow.
>>>>>>
>>>>>> David Ezzio
>>>>>>
>>>>>>
>>>>>> David Ezzio wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> I am convinced that change 616658 (small change to
>>>>>>> DBDictionary.java) is the source of the two intermittent
>>>>>>> failures
>>>> that I am seeing locally.
>>>>>>> One of those intermittent failures is the failure in
>>>>>>> TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
>>>>>>> which is one of the 33 related failures currently seen in
>>>>>>> TeamCity
>>>> Derby builds.
>>>>>>> The other intermittent failure is
>>>>>>> TestVersion.testVersionTimestamp
>>>>>>> which I reported earlier as occurring locally with change
>>> 616658.
>>>>>>> When I revert change 616658 locally, both of these issues go
>>>>>>> away even with the changes that I submitted to
>>>>>>> AbstractBrokerFactory in
>>>>>>> change 616972.
>>>>>>>
>>>>>>> I have a question: is it possible to run a build in TeamCity
>>>>>>> without submitting the change (or conditionally submitting the
>>>> change) to SVN?
>>>>>>> I'd like to find out if reverting 616658 will fix the issue. If
>>>>>>> so, then I can try to figure out why it's a problem.
>>>>>>>
>>>>>>> Thanks,
>>>>>>>
>>>>>>> David
>>>>>>>
>>>>>>> David Ezzio wrote:
>>>>>>>> No, but I'm taking a look since my changes appear to be the
>>>> cause.
>>>>>>>> Patrick Linskey wrote:
>>>>>>>>> 5. The tests are now failing in the automated build; any idea
>>>>>>>>> why this might be?
>>>>>>>>>
>>>>>>>>> -Patrick
>>>>>>>>>
>>>>>>>>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
>>>> wrote:
>>>>>>>>>> Hi,
>>>>>>>>>>
>>>>>>>>>> A few questions:
>>>>>>>>>>
>>>>>>>>>> 1. Were you seeing this cause problems somewhere? I.e., when
>>>>>>>>>> do
>>>>>>>>>> we serialize BrokerFactory instances?
>>>>>>>>>>
>>>>>>>>>> 2. Why not just move the initialization code to the
>>>>>>>>>> declaration
>>>>>>>>>> line (instead of the constructor)?
>>>>>>>>>>
>>>>>>>>>> 3. It looks like there are other transactional structures
>>>>>>>>>> that might not be properly handled in serialization /
>>>>>>>>>> deserialization (lifecycle listeners, for example). Should we
>>>> be making more changes there?
>>>>>>>>>> -Patrick
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
>>>>>>>>>>> Author: dezzio
>>>>>>>>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
>>>>>>>>>>>
>>>>>>>>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
>>>>>>>>>>> Log:
>>>>>>>>>>> Allow EntityManagerFactory objects to be serialized and
>>>>>>>>>>> deserialized successfully.
>>>>>>>>>>>
>>>>>>>>>>> Added:
>>>>>>>>>>>
>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>> (with props)
>>>>>>>>>>> Modified:
>>>>>>>>>>>
>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Modified:
>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>>>
>>>>>>>>>>> URL:
>>>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/sr
>>>>>>>>>>> c/
>>>>>>>>>>> main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.ja
>>>>>>>>>>> va ?rev=616972&r1=616971&r2=616972&view=diff
>>>>>>>>>>>
>>>>>>>>>>> ============================================================
>>>>>>>>>>> ==
>>>>>>>>>>> ================
>>>>>>>>>>>
>>>>>>>>>>> ---
>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>>> (original)
>>>>>>>>>>> +++
>>>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>>>> a/ kernel/AbstractBrokerFactory.java Wed Jan 30 16:59:02
>>>>>>>>>>> 2008 @@ -147,8 +147,7 @@
>>>>>>>>>>> */
>>>>>>>>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
>>>> config) {
>>>>>>>>>>> _conf = config;
>>>>>>>>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>>>>>>>> - ConcurrentReferenceHashSet.WEAK);
>>>>>>>>>>> + getPcClassLoaders();
>>>>>>>>>>> }
>>>>>>>>>>>
>>>>>>>>>>> /**
>>>>>>>>>>> @@ -287,13 +286,13 @@
>>>>>>>>>>> if (needsSub(cls))
>>>>>>>>>>> toRedefine.add(cls);
>>>>>>>>>>> }
>>>>>>>>>>> - _pcClassLoaders.add(loader);
>>>>>>>>>>> + getPcClassLoaders().add(loader);
>>>>>>>>>>> _pcClassNames = c;
>>>>>>>>>>> }
>>>>>>>>>>> _persistentTypesLoaded = true;
>>>>>>>>>>> } else {
>>>>>>>>>>> // reload with this loader
>>>>>>>>>>> - if (_pcClassLoaders.add(loader)) {
>>>>>>>>>>> + if (getPcClassLoaders().add(loader)) {
>>>>>>>>>>> for (Iterator itr =
>>>>>>>>>>> _pcClassNames.iterator();
>>>>>>>>>>> itr.hasNext();) {
>>>>>>>>>>> try {
>>>>>>>>>>> Class cls = @@ -818,4 +817,15 @@
>>>>>>>>>>> _transactional.remove (_trans);
>>>>>>>>>>> }
>>>>>>>>>>> }
>>>>>>>>>>> +
>>>>>>>>>>> + /**
>>>>>>>>>>> + * Method insures that deserialized EMF has this
>>>>>>>>>>> + reference
>>>>>>>>>>> re-instantiated
>>>>>>>>>>> + */
>>>>>>>>>>> + private Collection getPcClassLoaders() {
>>>>>>>>>>> + if (_pcClassLoaders == null)
>>>>>>>>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>>>>>>>> + ConcurrentReferenceHashSet.WEAK);
>>>>>>>>>>> +
>>>>>>>>>>> + return _pcClassLoaders;
>>>>>>>>>>> + }
>>>>>>>>>>> }
>>>>>>>>>>>
>>>>>>>>>>> Added:
>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>>
>>>>>>>>>>> URL:
>>>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persisten
>>>>>>>>>>> ce
>>>>>>>>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/Te
>>>>>>>>>>> st SerializedFactory.java?rev=616972&view=auto
>>>>>>>>>>>
>>>>>>>>>>> ============================================================
>>>>>>>>>>> ==
>>>>>>>>>>> ================
>>>>>>>>>>>
>>>>>>>>>>> ---
>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>> (added)
>>>>>>>>>>> +++
>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>> Wed Jan 30 16:59:02 2008
>>>>>>>>>>> @@ -0,0 +1,74 @@
>>>>>>>>>>> +/*
>>>>>>>>>>> + * Licensed to the Apache Software Foundation (ASF) under
>>>>>>>>>>> +one
>>>>>>>>>>> + * or more contributor license agreements. See the NOTICE
>>>>>>>>>>> +file
>>>>>>>>>>> + * distributed with this work for additional information
>>>>>>>>>>> + * regarding copyright ownership. The ASF licenses this
>>>>>>>>>>> +file
>>>>>>>>>>> + * to you 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 org.apache.openjpa.persistence.simple;
>>>>>>>>>>> +
>>>>>>>>>>> +import java.io.*;
>>>>>>>>>>> +
>>>>>>>>>>> +import javax.persistence.EntityManager; import
>>>>>>>>>>> +javax.persistence.EntityManagerFactory;
>>>>>>>>>>> +import javax.persistence.EntityTransaction;
>>>>>>>>>>> +
>>>>>>>>>>> +import junit.textui.TestRunner; import
>>>>>>>>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
>>>>>>>>>>> +import
>>>>>>>>>>> +org.apache.openjpa.persistence.test.SingleEMFTestCase;
>>>>>>>>>>> +
>>>>>>>>>>> +/**
>>>>>>>>>>> + * Tests that a EntityManagerFactory can be used after
>>>> serialization.
>>>>>>>>>>> + *
>>>>>>>>>>> + * @author David Ezzio
>>>>>>>>>>> + */
>>>>>>>>>>> +public class TestSerializedFactory
>>>>>>>>>>> + extends SingleEMFTestCase {
>>>>>>>>>>> +
>>>>>>>>>>> + public void setUp() {
>>>>>>>>>>> + setUp(AllFieldTypes.class);
>>>>>>>>>>> + }
>>>>>>>>>>> +
>>>>>>>>>>> + public void testSerializedEntityManagerFactory() throws
>>>>>>>>>>> Exception {
>>>>>>>>>>> + // serialize and deserialize the entity manager
>>>> factory
>>>>>>>>>>> + ByteArrayOutputStream baos = new
>>>> ByteArrayOutputStream();
>>>>>>>>>>> + ObjectOutputStream oos = new
>>>> ObjectOutputStream(baos);
>>>>>>>>>>> + oos.writeObject(emf);
>>>>>>>>>>> + EntityManagerFactory emf2 =
>>>>>>>>>>> + (EntityManagerFactory) new ObjectInputStream(
>>>>>>>>>>> + new
>>>>>>>>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
>>>>>>>>>>> +
>>>>>>>>>>> + // use the deserialized entity manager factory
>>>>>>>>>>> + assertTrue("The deserialized entity manager factory
>>>>>>>>>>> + is not
>>>>>>>>>>> open",
>>>>>>>>>>> + emf2.isOpen());
>>>>>>>>>>> + EntityManager em = emf2.createEntityManager();
>>>>>>>>>>> + assertTrue("The newly created entity manager is not
>>>>>>>>>>> + open",
>>>>>>>>>>> em.isOpen());
>>>>>>>>>>> +
>>>>>>>>>>> + // exercise the entity manager produced from the
>>>>>>>>>>> deserialized EMF
>>>>>>>>>>> + em.getTransaction().begin();
>>>>>>>>>>> + em.persist(new AllFieldTypes());
>>>>>>>>>>> + em.getTransaction().commit();
>>>>>>>>>>> +
>>>>>>>>>>> + // close the extra resources
>>>>>>>>>>> + em.close();
>>>>>>>>>>> + assertFalse("The entity manager is not closed",
>>>> em.isOpen());
>>>>>>>>>>> + emf2.close();
>>>>>>>>>>> + assertFalse("The entity manager factory is not
>>>>>>>>>>> + closed",
>>>>>>>>>>> emf2.isOpen());
>>>>>>>>>>> + }
>>>>>>>>>>> +
>>>>>>>>>>> + public static void main(String[] args) {
>>>>>>>>>>> + TestRunner.run(TestSerializedFactory.class);
>>>>>>>>>>> + }
>>>>>>>>>>> +}
>>>>>>>>>>> +
>>>>>>>>>>>
>>>>>>>>>>> Propchange:
>>>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>>>
>>>>>>>>>>> ------------------------------------------------------------
>>>>>>>>>>> --
>>>>>>>>>>> ----------------
>>>>>>>>>>>
>>>>>>>>>>> svn:eol-style = native
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>> Patrick Linskey
>>>>>>>>>> 202 669 5907
>>>>>>>>>>
>>>>>>>>>
>>>>>
>>>>> --
>>>>> Patrick Linskey
>>>>> 202 669 5907
>>>>>
>>>>
>>>> --
>>>> Patrick Linskey
>>>> 202 669 5907
>>>>
>>>> Notice: This email message, together with any attachments, may
>>>> contain information of BEA Systems, Inc., its subsidiaries and
>>>> affiliated entities, that may be confidential, proprietary,
>>>> copyrighted and/or legally privileged, and is intended solely for the
>>>> use of the individual or entity named in this message. If you are not
>>>> the intended recipient, and have received this message in error,
>>>> please immediately return this by email and then delete it.
>>>>
>>> Notice: This email message, together with any attachments, may contain
>> information of BEA Systems, Inc., its subsidiaries and affiliated
>> entities, that may be confidential, proprietary, copyrighted and/or
>> legally privileged, and is intended solely for the use of the individual or
>> entity named in this message. If you are not the intended recipient, and
>> have received this message in error, please immediately return this by email
>> and then delete it.
>
>
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by Patrick Linskey <pl...@gmail.com>.
Yes, I agree that if we make serialization not work, we should make
the BF not implement Serializable. That is my preferred approach: even
if Kodo needs to serialize a PMF for JNDI purposes or something, we
could do that in the PMF impl, instead of in the BF.
IMO, the build is fragile because it runs multiple unrelated tests in
the same VM with no strong isolation between tests. This leads to
unexpected side effects sometimes. This can be frustrating, but the
benefits are considerable: faster test run time, plus we get to find
out about weird unintended side effects.
-Patrick
On 2/1/08, David Ezzio <de...@apache.org> wrote:
> Hi,
>
> As to testing that the EMF objects are useful after serialization, this
> is a reasonable expectation of any object that is an instance of the
> Serializable interface.
>
> If it turns out that we don't want to support serializing useful EMF
> objects, then we should take care to remove the Serializable interface
> from their implementation.
>
> As I understand it, it is this latter issue that is currently under
> discussion.
>
> An important, but only tangentially related, issue is why the build is
> so fragile.
>
> David
>
> Pinaki Poddar wrote:
> > Hi,
> > Serializablity requirement on EMF is arising out of necessity rather
> > than by JPA specifcation.
> > JNDI, too, does not mandate bound objects to be Serializable.
> > But JNDI providers often place that restriction.
> > Non-serializable objects can be bound to JNDI by implementing
> > javax.naming.Referenceable.
> > Kodo's JDO product derivation based on OpenJPA is following this
> > Referenceable route.
> >
> > But I do not see this in openjpa EntityManagerFactory implementation.
> >
> >
> > -----Original Message-----
> > From: Kevin Sutter [mailto:kwsutter@gmail.com]
> > Sent: Friday, February 01, 2008 11:11 AM
> > To: dev@openjpa.apache.org
> > Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
> >
> > On Feb 1, 2008 10:17 AM, Pinaki Poddar <pp...@bea.com> wrote:
> >
> >> Factories needs to be Serializable as they are often placed in JNDI.
> >
> >
> > The JNDI lookup needs to work for the Factory definition, but it doesn't
> > require that the actual Factory be present in JNDI. And, since the JNDI
> > lookup is meant for the EE (Container) environment, I would guess that
> > most implementations provide a wrapper object in JNDI for the actual
> > Factory.
> >
> > Granted, this wrapper approach is not a requirement. But, I can't find
> > the requirement in the spec that the EMF has to be Serializable
> > either... :-)
> >
> > Kevin
> >
> >
> >> -----Original Message-----
> >> From: Patrick Linskey [mailto:plinskey@gmail.com]
> >> Sent: Thursday, January 31, 2008 7:57 PM
> >> To: dev@openjpa.apache.org
> >> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
> >>
> >> To be honest, I don't really understand why a BrokerFactory needs to
> >> be serializable, so I think probably it'd be good to make sure that
> >> there is good reason for that before going too deep into this
> >> debugging exercise.
> >>
> >> -Patrick
> >>
> >> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
> >>>> If the hangup is intolerable, let me know, and I'll back out my
> >>>> change tomorrow.
> >>> Personally, I'd rather have the TeamCity build be clean while you
> >>> investigate. As you noted, the error-reporting from surefire is
> >>> frustratingly inaccurate, and this really erodes the benefit of
> >>> continuous integration.
> >>>
> >>>> It would be handy if I knew a few things. How to pass in VM debug
> >
> >>>> args to Maven for its test runs, where to set the OpenJPA logging
> >>>> when Maven runs the tests, and how to pass in the appropriate
> >>>> parameters when running the test case outside of Maven. Any
> >>>> insights on these questions will be appreciated.
> >>> I usually just add log configuration to the test cases themselves,
> >>> in the setUp(...) method:
> >>>
> >>> setUp(Class.class, Class2.class, ..., "openjpa.Log",
> >>> "SQL=TRACE");
> >>>
> >>> I believe that it is also possible to set this at the command line
> >>> via
> >>> -Dopenjpa.Log=SQL=TRACE types of clauses.
> >>>
> >>> -Patrick
> >>>
> >>>
> >>> On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
> >>>> Hi,
> >>>>
> >>>> Progress has been slow on the test failures apparently introduced
> >>>> by
> >>>> 616972. I have made progress in reproducing the results.
> >>>>
> >>>> The miscounting in TeamCity is gross. Although it lists 33
> >>>> failing tests complete with a credible stack trace for each, in
> >>>> fact, there is only one failing test.
> >>>>
> >>>> Meanwhile, I believe now, that change 616658 is no longer a
> > suspect.
> >>>> I get very erratic results here, but they seem to vary more by
> >>>> whether I run "mvn test" after "mvn clean" or "mvn install" after
> >>>> "mvn clean". By doing just what TeamCity does, "mvn clean compile
> >
> >>>> test", I can finally reproduce the results locally that TeamCity
> >>>> reports in its Compile-Derby build for this change (395, I think).
> >>>> In spite of this, if I run "mvn
> >>>> -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
> >> test passes.
> >>>> It would be handy if I knew a few things. How to pass in VM debug
> >
> >>>> args to Maven for its test runs, where to set the OpenJPA logging
> >>>> when Maven runs the tests, and how to pass in the appropriate
> >>>> parameters when running the test case outside of Maven. Any
> >>>> insights on these questions will be appreciated.
> >>>>
> >>>> But at least at this point, I can reproduce locally what TeamCity
> >>>> reports, and although it happened only once out of one try, I am
> >>>> optimistic that it is reproducible. I will continue working this
> >>>> tomorrow.
> >>>>
> >>>> If the hangup is intolerable, let me know, and I'll back out my
> >>>> change tomorrow.
> >>>>
> >>>> David Ezzio
> >>>>
> >>>>
> >>>> David Ezzio wrote:
> >>>>> Hi,
> >>>>>
> >>>>> I am convinced that change 616658 (small change to
> >>>>> DBDictionary.java) is the source of the two intermittent
> >>>>> failures
> >> that I am seeing locally.
> >>>>> One of those intermittent failures is the failure in
> >>>>> TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
> >>>>> which is one of the 33 related failures currently seen in
> >>>>> TeamCity
> >> Derby builds.
> >>>>> The other intermittent failure is
> >>>>> TestVersion.testVersionTimestamp
> >>>>> which I reported earlier as occurring locally with change
> > 616658.
> >>>>> When I revert change 616658 locally, both of these issues go
> >>>>> away even with the changes that I submitted to
> >>>>> AbstractBrokerFactory in
> >>>>> change 616972.
> >>>>>
> >>>>> I have a question: is it possible to run a build in TeamCity
> >>>>> without submitting the change (or conditionally submitting the
> >> change) to SVN?
> >>>>> I'd like to find out if reverting 616658 will fix the issue. If
> >
> >>>>> so, then I can try to figure out why it's a problem.
> >>>>>
> >>>>> Thanks,
> >>>>>
> >>>>> David
> >>>>>
> >>>>> David Ezzio wrote:
> >>>>>> No, but I'm taking a look since my changes appear to be the
> >> cause.
> >>>>>> Patrick Linskey wrote:
> >>>>>>> 5. The tests are now failing in the automated build; any idea
> >>>>>>> why this might be?
> >>>>>>>
> >>>>>>> -Patrick
> >>>>>>>
> >>>>>>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
> >> wrote:
> >>>>>>>> Hi,
> >>>>>>>>
> >>>>>>>> A few questions:
> >>>>>>>>
> >>>>>>>> 1. Were you seeing this cause problems somewhere? I.e., when
> >>>>>>>> do
> >>>>>>>> we serialize BrokerFactory instances?
> >>>>>>>>
> >>>>>>>> 2. Why not just move the initialization code to the
> >>>>>>>> declaration
> >>>>>>>> line (instead of the constructor)?
> >>>>>>>>
> >>>>>>>> 3. It looks like there are other transactional structures
> >>>>>>>> that might not be properly handled in serialization /
> >>>>>>>> deserialization (lifecycle listeners, for example). Should we
> >> be making more changes there?
> >>>>>>>> -Patrick
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
> >>>>>>>>> Author: dezzio
> >>>>>>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
> >>>>>>>>>
> >>>>>>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
> >>>>>>>>> Log:
> >>>>>>>>> Allow EntityManagerFactory objects to be serialized and
> >>>>>>>>> deserialized successfully.
> >>>>>>>>>
> >>>>>>>>> Added:
> >>>>>>>>>
> >>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> >>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>> (with props)
> >>>>>>>>> Modified:
> >>>>>>>>>
> >>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
> >>>>>>>>> a/ kernel/AbstractBrokerFactory.java
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> Modified:
> >>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
> >>>>>>>>> a/ kernel/AbstractBrokerFactory.java
> >>>>>>>>>
> >>>>>>>>> URL:
> >>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/sr
> >>>>>>>>> c/
> >>>>>>>>> main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.ja
> >>>>>>>>> va ?rev=616972&r1=616971&r2=616972&view=diff
> >>>>>>>>>
> >>>>>>>>> ============================================================
> >>>>>>>>> ==
> >>>>>>>>> ================
> >>>>>>>>>
> >>>>>>>>> ---
> >>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
> >>>>>>>>> a/ kernel/AbstractBrokerFactory.java
> >>>>>>>>> (original)
> >>>>>>>>> +++
> >>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
> >>>>>>>>> a/ kernel/AbstractBrokerFactory.java Wed Jan 30 16:59:02
> >>>>>>>>> 2008 @@ -147,8 +147,7 @@
> >>>>>>>>> */
> >>>>>>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
> >> config) {
> >>>>>>>>> _conf = config;
> >>>>>>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
> >>>>>>>>> - ConcurrentReferenceHashSet.WEAK);
> >>>>>>>>> + getPcClassLoaders();
> >>>>>>>>> }
> >>>>>>>>>
> >>>>>>>>> /**
> >>>>>>>>> @@ -287,13 +286,13 @@
> >>>>>>>>> if (needsSub(cls))
> >>>>>>>>> toRedefine.add(cls);
> >>>>>>>>> }
> >>>>>>>>> - _pcClassLoaders.add(loader);
> >>>>>>>>> + getPcClassLoaders().add(loader);
> >>>>>>>>> _pcClassNames = c;
> >>>>>>>>> }
> >>>>>>>>> _persistentTypesLoaded = true;
> >>>>>>>>> } else {
> >>>>>>>>> // reload with this loader
> >>>>>>>>> - if (_pcClassLoaders.add(loader)) {
> >>>>>>>>> + if (getPcClassLoaders().add(loader)) {
> >>>>>>>>> for (Iterator itr =
> >>>>>>>>> _pcClassNames.iterator();
> >>>>>>>>> itr.hasNext();) {
> >>>>>>>>> try {
> >>>>>>>>> Class cls = @@ -818,4 +817,15 @@
> >>>>>>>>> _transactional.remove (_trans);
> >>>>>>>>> }
> >>>>>>>>> }
> >>>>>>>>> +
> >>>>>>>>> + /**
> >>>>>>>>> + * Method insures that deserialized EMF has this
> >>>>>>>>> + reference
> >>>>>>>>> re-instantiated
> >>>>>>>>> + */
> >>>>>>>>> + private Collection getPcClassLoaders() {
> >>>>>>>>> + if (_pcClassLoaders == null)
> >>>>>>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
> >>>>>>>>> + ConcurrentReferenceHashSet.WEAK);
> >>>>>>>>> +
> >>>>>>>>> + return _pcClassLoaders;
> >>>>>>>>> + }
> >>>>>>>>> }
> >>>>>>>>>
> >>>>>>>>> Added:
> >>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> >>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>>
> >>>>>>>>> URL:
> >>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persisten
> >>>>>>>>> ce
> >>>>>>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/Te
> >>>>>>>>> st SerializedFactory.java?rev=616972&view=auto
> >>>>>>>>>
> >>>>>>>>> ============================================================
> >>>>>>>>> ==
> >>>>>>>>> ================
> >>>>>>>>>
> >>>>>>>>> ---
> >>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> >>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>> (added)
> >>>>>>>>> +++
> >>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> >>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>> Wed Jan 30 16:59:02 2008
> >>>>>>>>> @@ -0,0 +1,74 @@
> >>>>>>>>> +/*
> >>>>>>>>> + * Licensed to the Apache Software Foundation (ASF) under
> >>>>>>>>> +one
> >>>>>>>>> + * or more contributor license agreements. See the NOTICE
> >>>>>>>>> +file
> >>>>>>>>> + * distributed with this work for additional information
> >>>>>>>>> + * regarding copyright ownership. The ASF licenses this
> >>>>>>>>> +file
> >>>>>>>>> + * to you 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 org.apache.openjpa.persistence.simple;
> >>>>>>>>> +
> >>>>>>>>> +import java.io.*;
> >>>>>>>>> +
> >>>>>>>>> +import javax.persistence.EntityManager; import
> >>>>>>>>> +javax.persistence.EntityManagerFactory;
> >>>>>>>>> +import javax.persistence.EntityTransaction;
> >>>>>>>>> +
> >>>>>>>>> +import junit.textui.TestRunner; import
> >>>>>>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
> >>>>>>>>> +import
> >>>>>>>>> +org.apache.openjpa.persistence.test.SingleEMFTestCase;
> >>>>>>>>> +
> >>>>>>>>> +/**
> >>>>>>>>> + * Tests that a EntityManagerFactory can be used after
> >> serialization.
> >>>>>>>>> + *
> >>>>>>>>> + * @author David Ezzio
> >>>>>>>>> + */
> >>>>>>>>> +public class TestSerializedFactory
> >>>>>>>>> + extends SingleEMFTestCase {
> >>>>>>>>> +
> >>>>>>>>> + public void setUp() {
> >>>>>>>>> + setUp(AllFieldTypes.class);
> >>>>>>>>> + }
> >>>>>>>>> +
> >>>>>>>>> + public void testSerializedEntityManagerFactory() throws
> >>>>>>>>> Exception {
> >>>>>>>>> + // serialize and deserialize the entity manager
> >> factory
> >>>>>>>>> + ByteArrayOutputStream baos = new
> >> ByteArrayOutputStream();
> >>>>>>>>> + ObjectOutputStream oos = new
> >> ObjectOutputStream(baos);
> >>>>>>>>> + oos.writeObject(emf);
> >>>>>>>>> + EntityManagerFactory emf2 =
> >>>>>>>>> + (EntityManagerFactory) new ObjectInputStream(
> >>>>>>>>> + new
> >>>>>>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
> >>>>>>>>> +
> >>>>>>>>> + // use the deserialized entity manager factory
> >>>>>>>>> + assertTrue("The deserialized entity manager factory
> >
> >>>>>>>>> + is not
> >>>>>>>>> open",
> >>>>>>>>> + emf2.isOpen());
> >>>>>>>>> + EntityManager em = emf2.createEntityManager();
> >>>>>>>>> + assertTrue("The newly created entity manager is not
> >
> >>>>>>>>> + open",
> >>>>>>>>> em.isOpen());
> >>>>>>>>> +
> >>>>>>>>> + // exercise the entity manager produced from the
> >>>>>>>>> deserialized EMF
> >>>>>>>>> + em.getTransaction().begin();
> >>>>>>>>> + em.persist(new AllFieldTypes());
> >>>>>>>>> + em.getTransaction().commit();
> >>>>>>>>> +
> >>>>>>>>> + // close the extra resources
> >>>>>>>>> + em.close();
> >>>>>>>>> + assertFalse("The entity manager is not closed",
> >> em.isOpen());
> >>>>>>>>> + emf2.close();
> >>>>>>>>> + assertFalse("The entity manager factory is not
> >>>>>>>>> + closed",
> >>>>>>>>> emf2.isOpen());
> >>>>>>>>> + }
> >>>>>>>>> +
> >>>>>>>>> + public static void main(String[] args) {
> >>>>>>>>> + TestRunner.run(TestSerializedFactory.class);
> >>>>>>>>> + }
> >>>>>>>>> +}
> >>>>>>>>> +
> >>>>>>>>>
> >>>>>>>>> Propchange:
> >>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> >>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>>>>>
> >>>>>>>>> ------------------------------------------------------------
> >>>>>>>>> --
> >>>>>>>>> ----------------
> >>>>>>>>>
> >>>>>>>>> svn:eol-style = native
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>
> >>>>>>>> --
> >>>>>>>> Patrick Linskey
> >>>>>>>> 202 669 5907
> >>>>>>>>
> >>>>>>>
> >>>>>>>
> >>>
> >>>
> >>> --
> >>> Patrick Linskey
> >>> 202 669 5907
> >>>
> >>
> >>
> >> --
> >> Patrick Linskey
> >> 202 669 5907
> >>
> >> Notice: This email message, together with any attachments, may
> >> contain information of BEA Systems, Inc., its subsidiaries and
> >> affiliated entities, that may be confidential, proprietary,
> >> copyrighted and/or legally privileged, and is intended solely for the
> >
> >> use of the individual or entity named in this message. If you are not
> >> the intended recipient, and have received this message in error,
> >> please immediately return this by email and then delete it.
> >>
> >
> > Notice: This email message, together with any attachments, may contain
> information of BEA Systems, Inc., its subsidiaries and affiliated
> entities, that may be confidential, proprietary, copyrighted and/or
> legally privileged, and is intended solely for the use of the individual or
> entity named in this message. If you are not the intended recipient, and
> have received this message in error, please immediately return this by email
> and then delete it.
> >
>
--
Patrick Linskey
202 669 5907
Does EMF need to be Serializable?
Posted by Pinaki Poddar <pp...@bea.com>.
Hi,
If serializability is only required because EMF need to be bound to
JNDI context, then it may sense to make EMFImpl implement
javax.naming.Referenceable and no more java.io.Serializable.
This may be the preferred route because EMF configuration is now partly
dynamic. If EMF is bound to JNDI as a serialized instance, then any
dynamic change to its configuration should rebind the EMF to JNDI.
-----Original Message-----
From: David Ezzio [mailto:dezzio@apache.org]
Sent: Friday, February 01, 2008 12:11 PM
To: dev@openjpa.apache.org
Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
Hi,
As to testing that the EMF objects are useful after serialization, this
is a reasonable expectation of any object that is an instance of the
Serializable interface.
If it turns out that we don't want to support serializing useful EMF
objects, then we should take care to remove the Serializable interface
from their implementation.
As I understand it, it is this latter issue that is currently under
discussion.
An important, but only tangentially related, issue is why the build is
so fragile.
David
Pinaki Poddar wrote:
> Hi,
> Serializablity requirement on EMF is arising out of necessity rather
> than by JPA specifcation.
> JNDI, too, does not mandate bound objects to be Serializable.
> But JNDI providers often place that restriction.
> Non-serializable objects can be bound to JNDI by implementing
> javax.naming.Referenceable.
> Kodo's JDO product derivation based on OpenJPA is following this
> Referenceable route.
>
> But I do not see this in openjpa EntityManagerFactory implementation.
>
>
> -----Original Message-----
> From: Kevin Sutter [mailto:kwsutter@gmail.com]
> Sent: Friday, February 01, 2008 11:11 AM
> To: dev@openjpa.apache.org
> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
>
> On Feb 1, 2008 10:17 AM, Pinaki Poddar <pp...@bea.com> wrote:
>
>> Factories needs to be Serializable as they are often placed in JNDI.
>
>
> The JNDI lookup needs to work for the Factory definition, but it
> doesn't require that the actual Factory be present in JNDI. And,
> since the JNDI lookup is meant for the EE (Container) environment, I
> would guess that most implementations provide a wrapper object in JNDI
> for the actual Factory.
>
> Granted, this wrapper approach is not a requirement. But, I can't
> find the requirement in the spec that the EMF has to be Serializable
> either... :-)
>
> Kevin
>
>
>> -----Original Message-----
>> From: Patrick Linskey [mailto:plinskey@gmail.com]
>> Sent: Thursday, January 31, 2008 7:57 PM
>> To: dev@openjpa.apache.org
>> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
>>
>> To be honest, I don't really understand why a BrokerFactory needs to
>> be serializable, so I think probably it'd be good to make sure that
>> there is good reason for that before going too deep into this
>> debugging exercise.
>>
>> -Patrick
>>
>> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
>>>> If the hangup is intolerable, let me know, and I'll back out my
>>>> change tomorrow.
>>> Personally, I'd rather have the TeamCity build be clean while you
>>> investigate. As you noted, the error-reporting from surefire is
>>> frustratingly inaccurate, and this really erodes the benefit of
>>> continuous integration.
>>>
>>>> It would be handy if I knew a few things. How to pass in VM debug
>
>>>> args to Maven for its test runs, where to set the OpenJPA logging
>>>> when Maven runs the tests, and how to pass in the appropriate
>>>> parameters when running the test case outside of Maven. Any
>>>> insights on these questions will be appreciated.
>>> I usually just add log configuration to the test cases themselves,
>>> in the setUp(...) method:
>>>
>>> setUp(Class.class, Class2.class, ..., "openjpa.Log",
>>> "SQL=TRACE");
>>>
>>> I believe that it is also possible to set this at the command line
>>> via -Dopenjpa.Log=SQL=TRACE types of clauses.
>>>
>>> -Patrick
>>>
>>>
>>> On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
>>>> Hi,
>>>>
>>>> Progress has been slow on the test failures apparently introduced
>>>> by 616972. I have made progress in reproducing the results.
>>>>
>>>> The miscounting in TeamCity is gross. Although it lists 33 failing
>>>> tests complete with a credible stack trace for each, in fact, there
>>>> is only one failing test.
>>>>
>>>> Meanwhile, I believe now, that change 616658 is no longer a
> suspect.
>>>> I get very erratic results here, but they seem to vary more by
>>>> whether I run "mvn test" after "mvn clean" or "mvn install" after
>>>> "mvn clean". By doing just what TeamCity does, "mvn clean compile
>
>>>> test", I can finally reproduce the results locally that TeamCity
>>>> reports in its Compile-Derby build for this change (395, I think).
>>>> In spite of this, if I run "mvn
>>>> -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
>> test passes.
>>>> It would be handy if I knew a few things. How to pass in VM debug
>
>>>> args to Maven for its test runs, where to set the OpenJPA logging
>>>> when Maven runs the tests, and how to pass in the appropriate
>>>> parameters when running the test case outside of Maven. Any
>>>> insights on these questions will be appreciated.
>>>>
>>>> But at least at this point, I can reproduce locally what TeamCity
>>>> reports, and although it happened only once out of one try, I am
>>>> optimistic that it is reproducible. I will continue working this
>>>> tomorrow.
>>>>
>>>> If the hangup is intolerable, let me know, and I'll back out my
>>>> change tomorrow.
>>>>
>>>> David Ezzio
>>>>
>>>>
>>>> David Ezzio wrote:
>>>>> Hi,
>>>>>
>>>>> I am convinced that change 616658 (small change to
>>>>> DBDictionary.java) is the source of the two intermittent failures
>> that I am seeing locally.
>>>>> One of those intermittent failures is the failure in
>>>>> TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
>>>>> which is one of the 33 related failures currently seen in TeamCity
>> Derby builds.
>>>>> The other intermittent failure is TestVersion.testVersionTimestamp
>>>>> which I reported earlier as occurring locally with change
> 616658.
>>>>> When I revert change 616658 locally, both of these issues go away
>>>>> even with the changes that I submitted to AbstractBrokerFactory in
>>>>> change 616972.
>>>>>
>>>>> I have a question: is it possible to run a build in TeamCity
>>>>> without submitting the change (or conditionally submitting the
>> change) to SVN?
>>>>> I'd like to find out if reverting 616658 will fix the issue. If
>
>>>>> so, then I can try to figure out why it's a problem.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> David
>>>>>
>>>>> David Ezzio wrote:
>>>>>> No, but I'm taking a look since my changes appear to be the
>> cause.
>>>>>> Patrick Linskey wrote:
>>>>>>> 5. The tests are now failing in the automated build; any idea
>>>>>>> why this might be?
>>>>>>>
>>>>>>> -Patrick
>>>>>>>
>>>>>>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
>> wrote:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> A few questions:
>>>>>>>>
>>>>>>>> 1. Were you seeing this cause problems somewhere? I.e., when do
>>>>>>>> we serialize BrokerFactory instances?
>>>>>>>>
>>>>>>>> 2. Why not just move the initialization code to the declaration
>>>>>>>> line (instead of the constructor)?
>>>>>>>>
>>>>>>>> 3. It looks like there are other transactional structures that
>>>>>>>> might not be properly handled in serialization /
>>>>>>>> deserialization (lifecycle listeners, for example). Should we
>> be making more changes there?
>>>>>>>> -Patrick
>>>>>>>>
>>>>>>>>
>>>>>>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
>>>>>>>>> Author: dezzio
>>>>>>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
>>>>>>>>>
>>>>>>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
>>>>>>>>> Log:
>>>>>>>>> Allow EntityManagerFactory objects to be serialized and
>>>>>>>>> deserialized successfully.
>>>>>>>>>
>>>>>>>>> Added:
>>>>>>>>>
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>> (with props)
>>>>>>>>> Modified:
>>>>>>>>>
>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Modified:
>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>
>>>>>>>>> URL:
>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/sr
>>>>>>>>> c/
>>>>>>>>> main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.ja
>>>>>>>>> va ?rev=616972&r1=616971&r2=616972&view=diff
>>>>>>>>>
>>>>>>>>> ============================================================
>>>>>>>>> ==
>>>>>>>>> ================
>>>>>>>>>
>>>>>>>>> ---
>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>> (original)
>>>>>>>>> +++
>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>> a/ kernel/AbstractBrokerFactory.java Wed Jan 30 16:59:02
>>>>>>>>> 2008 @@ -147,8 +147,7 @@
>>>>>>>>> */
>>>>>>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
>> config) {
>>>>>>>>> _conf = config;
>>>>>>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>>>>>> - ConcurrentReferenceHashSet.WEAK);
>>>>>>>>> + getPcClassLoaders();
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> /**
>>>>>>>>> @@ -287,13 +286,13 @@
>>>>>>>>> if (needsSub(cls))
>>>>>>>>> toRedefine.add(cls);
>>>>>>>>> }
>>>>>>>>> - _pcClassLoaders.add(loader);
>>>>>>>>> + getPcClassLoaders().add(loader);
>>>>>>>>> _pcClassNames = c;
>>>>>>>>> }
>>>>>>>>> _persistentTypesLoaded = true;
>>>>>>>>> } else {
>>>>>>>>> // reload with this loader
>>>>>>>>> - if (_pcClassLoaders.add(loader)) {
>>>>>>>>> + if (getPcClassLoaders().add(loader)) {
>>>>>>>>> for (Iterator itr = _pcClassNames.iterator();
>>>>>>>>> itr.hasNext();) {
>>>>>>>>> try {
>>>>>>>>> Class cls = @@ -818,4 +817,15 @@
>>>>>>>>> _transactional.remove (_trans);
>>>>>>>>> }
>>>>>>>>> }
>>>>>>>>> +
>>>>>>>>> + /**
>>>>>>>>> + * Method insures that deserialized EMF has this reference
>>>>>>>>> re-instantiated
>>>>>>>>> + */
>>>>>>>>> + private Collection getPcClassLoaders() {
>>>>>>>>> + if (_pcClassLoaders == null)
>>>>>>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>>>>>> + ConcurrentReferenceHashSet.WEAK);
>>>>>>>>> +
>>>>>>>>> + return _pcClassLoaders;
>>>>>>>>> + }
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> Added:
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>
>>>>>>>>> URL:
>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persisten
>>>>>>>>> ce
>>>>>>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/Te
>>>>>>>>> st SerializedFactory.java?rev=616972&view=auto
>>>>>>>>>
>>>>>>>>> ============================================================
>>>>>>>>> ==
>>>>>>>>> ================
>>>>>>>>>
>>>>>>>>> ---
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>> (added)
>>>>>>>>> +++
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>> Wed Jan 30 16:59:02 2008
>>>>>>>>> @@ -0,0 +1,74 @@
>>>>>>>>> +/*
>>>>>>>>> + * Licensed to the Apache Software Foundation (ASF) under one
>>>>>>>>> + * or more contributor license agreements. See the NOTICE
>>>>>>>>> +file
>>>>>>>>> + * distributed with this work for additional information
>>>>>>>>> + * regarding copyright ownership. The ASF licenses this file
>>>>>>>>> + * to you 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 org.apache.openjpa.persistence.simple;
>>>>>>>>> +
>>>>>>>>> +import java.io.*;
>>>>>>>>> +
>>>>>>>>> +import javax.persistence.EntityManager; import
>>>>>>>>> +javax.persistence.EntityManagerFactory;
>>>>>>>>> +import javax.persistence.EntityTransaction;
>>>>>>>>> +
>>>>>>>>> +import junit.textui.TestRunner; import
>>>>>>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
>>>>>>>>> +import
>>>>>>>>> +org.apache.openjpa.persistence.test.SingleEMFTestCase;
>>>>>>>>> +
>>>>>>>>> +/**
>>>>>>>>> + * Tests that a EntityManagerFactory can be used after
>> serialization.
>>>>>>>>> + *
>>>>>>>>> + * @author David Ezzio
>>>>>>>>> + */
>>>>>>>>> +public class TestSerializedFactory
>>>>>>>>> + extends SingleEMFTestCase {
>>>>>>>>> +
>>>>>>>>> + public void setUp() {
>>>>>>>>> + setUp(AllFieldTypes.class);
>>>>>>>>> + }
>>>>>>>>> +
>>>>>>>>> + public void testSerializedEntityManagerFactory() throws
>>>>>>>>> Exception {
>>>>>>>>> + // serialize and deserialize the entity manager
>> factory
>>>>>>>>> + ByteArrayOutputStream baos = new
>> ByteArrayOutputStream();
>>>>>>>>> + ObjectOutputStream oos = new
>> ObjectOutputStream(baos);
>>>>>>>>> + oos.writeObject(emf);
>>>>>>>>> + EntityManagerFactory emf2 =
>>>>>>>>> + (EntityManagerFactory) new ObjectInputStream(
>>>>>>>>> + new
>>>>>>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
>>>>>>>>> +
>>>>>>>>> + // use the deserialized entity manager factory
>>>>>>>>> + assertTrue("The deserialized entity manager factory
>
>>>>>>>>> + is not
>>>>>>>>> open",
>>>>>>>>> + emf2.isOpen());
>>>>>>>>> + EntityManager em = emf2.createEntityManager();
>>>>>>>>> + assertTrue("The newly created entity manager is not
>
>>>>>>>>> + open",
>>>>>>>>> em.isOpen());
>>>>>>>>> +
>>>>>>>>> + // exercise the entity manager produced from the
>>>>>>>>> deserialized EMF
>>>>>>>>> + em.getTransaction().begin();
>>>>>>>>> + em.persist(new AllFieldTypes());
>>>>>>>>> + em.getTransaction().commit();
>>>>>>>>> +
>>>>>>>>> + // close the extra resources
>>>>>>>>> + em.close();
>>>>>>>>> + assertFalse("The entity manager is not closed",
>> em.isOpen());
>>>>>>>>> + emf2.close();
>>>>>>>>> + assertFalse("The entity manager factory is not
>>>>>>>>> + closed",
>>>>>>>>> emf2.isOpen());
>>>>>>>>> + }
>>>>>>>>> +
>>>>>>>>> + public static void main(String[] args) {
>>>>>>>>> + TestRunner.run(TestSerializedFactory.class);
>>>>>>>>> + }
>>>>>>>>> +}
>>>>>>>>> +
>>>>>>>>>
>>>>>>>>> Propchange:
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------
>>>>>>>>> --
>>>>>>>>> ----------------
>>>>>>>>>
>>>>>>>>> svn:eol-style = native
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Patrick Linskey
>>>>>>>> 202 669 5907
>>>>>>>>
>>>>>>>
>>>>>>>
>>>
>>>
>>> --
>>> Patrick Linskey
>>> 202 669 5907
>>>
>>
>>
>> --
>> Patrick Linskey
>> 202 669 5907
>>
>> Notice: This email message, together with any attachments, may
>> contain information of BEA Systems, Inc., its subsidiaries and
>> affiliated entities, that may be confidential, proprietary,
>> copyrighted and/or legally privileged, and is intended solely for
>> the
>
>> use of the individual or entity named in this message. If you are not
>> the intended recipient, and have received this message in error,
>> please immediately return this by email and then delete it.
>>
>
> Notice: This email message, together with any attachments, may
contain information of BEA Systems, Inc., its subsidiaries and
affiliated entities, that may be confidential, proprietary,
copyrighted and/or legally privileged, and is intended solely for the
use of the individual or entity named in this message. If you are not
the intended recipient, and have received this message in error, please
immediately return this by email and then delete it.
>
Notice: This email message, together with any attachments, may contain information of BEA Systems, Inc., its subsidiaries and affiliated entities, that may be confidential, proprietary, copyrighted and/or legally privileged, and is intended solely for the use of the individual or entity named in this message. If you are not the intended recipient, and have received this message in error, please immediately return this by email and then delete it.
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by David Ezzio <de...@apache.org>.
Hi,
As to testing that the EMF objects are useful after serialization, this
is a reasonable expectation of any object that is an instance of the
Serializable interface.
If it turns out that we don't want to support serializing useful EMF
objects, then we should take care to remove the Serializable interface
from their implementation.
As I understand it, it is this latter issue that is currently under
discussion.
An important, but only tangentially related, issue is why the build is
so fragile.
David
Pinaki Poddar wrote:
> Hi,
> Serializablity requirement on EMF is arising out of necessity rather
> than by JPA specifcation.
> JNDI, too, does not mandate bound objects to be Serializable.
> But JNDI providers often place that restriction.
> Non-serializable objects can be bound to JNDI by implementing
> javax.naming.Referenceable.
> Kodo's JDO product derivation based on OpenJPA is following this
> Referenceable route.
>
> But I do not see this in openjpa EntityManagerFactory implementation.
>
>
> -----Original Message-----
> From: Kevin Sutter [mailto:kwsutter@gmail.com]
> Sent: Friday, February 01, 2008 11:11 AM
> To: dev@openjpa.apache.org
> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
>
> On Feb 1, 2008 10:17 AM, Pinaki Poddar <pp...@bea.com> wrote:
>
>> Factories needs to be Serializable as they are often placed in JNDI.
>
>
> The JNDI lookup needs to work for the Factory definition, but it doesn't
> require that the actual Factory be present in JNDI. And, since the JNDI
> lookup is meant for the EE (Container) environment, I would guess that
> most implementations provide a wrapper object in JNDI for the actual
> Factory.
>
> Granted, this wrapper approach is not a requirement. But, I can't find
> the requirement in the spec that the EMF has to be Serializable
> either... :-)
>
> Kevin
>
>
>> -----Original Message-----
>> From: Patrick Linskey [mailto:plinskey@gmail.com]
>> Sent: Thursday, January 31, 2008 7:57 PM
>> To: dev@openjpa.apache.org
>> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
>>
>> To be honest, I don't really understand why a BrokerFactory needs to
>> be serializable, so I think probably it'd be good to make sure that
>> there is good reason for that before going too deep into this
>> debugging exercise.
>>
>> -Patrick
>>
>> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
>>>> If the hangup is intolerable, let me know, and I'll back out my
>>>> change tomorrow.
>>> Personally, I'd rather have the TeamCity build be clean while you
>>> investigate. As you noted, the error-reporting from surefire is
>>> frustratingly inaccurate, and this really erodes the benefit of
>>> continuous integration.
>>>
>>>> It would be handy if I knew a few things. How to pass in VM debug
>
>>>> args to Maven for its test runs, where to set the OpenJPA logging
>>>> when Maven runs the tests, and how to pass in the appropriate
>>>> parameters when running the test case outside of Maven. Any
>>>> insights on these questions will be appreciated.
>>> I usually just add log configuration to the test cases themselves,
>>> in the setUp(...) method:
>>>
>>> setUp(Class.class, Class2.class, ..., "openjpa.Log",
>>> "SQL=TRACE");
>>>
>>> I believe that it is also possible to set this at the command line
>>> via
>>> -Dopenjpa.Log=SQL=TRACE types of clauses.
>>>
>>> -Patrick
>>>
>>>
>>> On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
>>>> Hi,
>>>>
>>>> Progress has been slow on the test failures apparently introduced
>>>> by
>>>> 616972. I have made progress in reproducing the results.
>>>>
>>>> The miscounting in TeamCity is gross. Although it lists 33
>>>> failing tests complete with a credible stack trace for each, in
>>>> fact, there is only one failing test.
>>>>
>>>> Meanwhile, I believe now, that change 616658 is no longer a
> suspect.
>>>> I get very erratic results here, but they seem to vary more by
>>>> whether I run "mvn test" after "mvn clean" or "mvn install" after
>>>> "mvn clean". By doing just what TeamCity does, "mvn clean compile
>
>>>> test", I can finally reproduce the results locally that TeamCity
>>>> reports in its Compile-Derby build for this change (395, I think).
>>>> In spite of this, if I run "mvn
>>>> -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
>> test passes.
>>>> It would be handy if I knew a few things. How to pass in VM debug
>
>>>> args to Maven for its test runs, where to set the OpenJPA logging
>>>> when Maven runs the tests, and how to pass in the appropriate
>>>> parameters when running the test case outside of Maven. Any
>>>> insights on these questions will be appreciated.
>>>>
>>>> But at least at this point, I can reproduce locally what TeamCity
>>>> reports, and although it happened only once out of one try, I am
>>>> optimistic that it is reproducible. I will continue working this
>>>> tomorrow.
>>>>
>>>> If the hangup is intolerable, let me know, and I'll back out my
>>>> change tomorrow.
>>>>
>>>> David Ezzio
>>>>
>>>>
>>>> David Ezzio wrote:
>>>>> Hi,
>>>>>
>>>>> I am convinced that change 616658 (small change to
>>>>> DBDictionary.java) is the source of the two intermittent
>>>>> failures
>> that I am seeing locally.
>>>>> One of those intermittent failures is the failure in
>>>>> TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
>>>>> which is one of the 33 related failures currently seen in
>>>>> TeamCity
>> Derby builds.
>>>>> The other intermittent failure is
>>>>> TestVersion.testVersionTimestamp
>>>>> which I reported earlier as occurring locally with change
> 616658.
>>>>> When I revert change 616658 locally, both of these issues go
>>>>> away even with the changes that I submitted to
>>>>> AbstractBrokerFactory in
>>>>> change 616972.
>>>>>
>>>>> I have a question: is it possible to run a build in TeamCity
>>>>> without submitting the change (or conditionally submitting the
>> change) to SVN?
>>>>> I'd like to find out if reverting 616658 will fix the issue. If
>
>>>>> so, then I can try to figure out why it's a problem.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> David
>>>>>
>>>>> David Ezzio wrote:
>>>>>> No, but I'm taking a look since my changes appear to be the
>> cause.
>>>>>> Patrick Linskey wrote:
>>>>>>> 5. The tests are now failing in the automated build; any idea
>>>>>>> why this might be?
>>>>>>>
>>>>>>> -Patrick
>>>>>>>
>>>>>>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
>> wrote:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> A few questions:
>>>>>>>>
>>>>>>>> 1. Were you seeing this cause problems somewhere? I.e., when
>>>>>>>> do
>>>>>>>> we serialize BrokerFactory instances?
>>>>>>>>
>>>>>>>> 2. Why not just move the initialization code to the
>>>>>>>> declaration
>>>>>>>> line (instead of the constructor)?
>>>>>>>>
>>>>>>>> 3. It looks like there are other transactional structures
>>>>>>>> that might not be properly handled in serialization /
>>>>>>>> deserialization (lifecycle listeners, for example). Should we
>> be making more changes there?
>>>>>>>> -Patrick
>>>>>>>>
>>>>>>>>
>>>>>>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
>>>>>>>>> Author: dezzio
>>>>>>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
>>>>>>>>>
>>>>>>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
>>>>>>>>> Log:
>>>>>>>>> Allow EntityManagerFactory objects to be serialized and
>>>>>>>>> deserialized successfully.
>>>>>>>>>
>>>>>>>>> Added:
>>>>>>>>>
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>> (with props)
>>>>>>>>> Modified:
>>>>>>>>>
>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Modified:
>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>>
>>>>>>>>> URL:
>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/sr
>>>>>>>>> c/
>>>>>>>>> main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.ja
>>>>>>>>> va ?rev=616972&r1=616971&r2=616972&view=diff
>>>>>>>>>
>>>>>>>>> ============================================================
>>>>>>>>> ==
>>>>>>>>> ================
>>>>>>>>>
>>>>>>>>> ---
>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>> a/ kernel/AbstractBrokerFactory.java
>>>>>>>>> (original)
>>>>>>>>> +++
>>>>>>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
>>>>>>>>> a/ kernel/AbstractBrokerFactory.java Wed Jan 30 16:59:02
>>>>>>>>> 2008 @@ -147,8 +147,7 @@
>>>>>>>>> */
>>>>>>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
>> config) {
>>>>>>>>> _conf = config;
>>>>>>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>>>>>> - ConcurrentReferenceHashSet.WEAK);
>>>>>>>>> + getPcClassLoaders();
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> /**
>>>>>>>>> @@ -287,13 +286,13 @@
>>>>>>>>> if (needsSub(cls))
>>>>>>>>> toRedefine.add(cls);
>>>>>>>>> }
>>>>>>>>> - _pcClassLoaders.add(loader);
>>>>>>>>> + getPcClassLoaders().add(loader);
>>>>>>>>> _pcClassNames = c;
>>>>>>>>> }
>>>>>>>>> _persistentTypesLoaded = true;
>>>>>>>>> } else {
>>>>>>>>> // reload with this loader
>>>>>>>>> - if (_pcClassLoaders.add(loader)) {
>>>>>>>>> + if (getPcClassLoaders().add(loader)) {
>>>>>>>>> for (Iterator itr =
>>>>>>>>> _pcClassNames.iterator();
>>>>>>>>> itr.hasNext();) {
>>>>>>>>> try {
>>>>>>>>> Class cls = @@ -818,4 +817,15 @@
>>>>>>>>> _transactional.remove (_trans);
>>>>>>>>> }
>>>>>>>>> }
>>>>>>>>> +
>>>>>>>>> + /**
>>>>>>>>> + * Method insures that deserialized EMF has this
>>>>>>>>> + reference
>>>>>>>>> re-instantiated
>>>>>>>>> + */
>>>>>>>>> + private Collection getPcClassLoaders() {
>>>>>>>>> + if (_pcClassLoaders == null)
>>>>>>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
>>>>>>>>> + ConcurrentReferenceHashSet.WEAK);
>>>>>>>>> +
>>>>>>>>> + return _pcClassLoaders;
>>>>>>>>> + }
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> Added:
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>
>>>>>>>>> URL:
>>>>>>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persisten
>>>>>>>>> ce
>>>>>>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/Te
>>>>>>>>> st SerializedFactory.java?rev=616972&view=auto
>>>>>>>>>
>>>>>>>>> ============================================================
>>>>>>>>> ==
>>>>>>>>> ================
>>>>>>>>>
>>>>>>>>> ---
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>> (added)
>>>>>>>>> +++
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>> Wed Jan 30 16:59:02 2008
>>>>>>>>> @@ -0,0 +1,74 @@
>>>>>>>>> +/*
>>>>>>>>> + * Licensed to the Apache Software Foundation (ASF) under
>>>>>>>>> +one
>>>>>>>>> + * or more contributor license agreements. See the NOTICE
>>>>>>>>> +file
>>>>>>>>> + * distributed with this work for additional information
>>>>>>>>> + * regarding copyright ownership. The ASF licenses this
>>>>>>>>> +file
>>>>>>>>> + * to you 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 org.apache.openjpa.persistence.simple;
>>>>>>>>> +
>>>>>>>>> +import java.io.*;
>>>>>>>>> +
>>>>>>>>> +import javax.persistence.EntityManager; import
>>>>>>>>> +javax.persistence.EntityManagerFactory;
>>>>>>>>> +import javax.persistence.EntityTransaction;
>>>>>>>>> +
>>>>>>>>> +import junit.textui.TestRunner; import
>>>>>>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
>>>>>>>>> +import
>>>>>>>>> +org.apache.openjpa.persistence.test.SingleEMFTestCase;
>>>>>>>>> +
>>>>>>>>> +/**
>>>>>>>>> + * Tests that a EntityManagerFactory can be used after
>> serialization.
>>>>>>>>> + *
>>>>>>>>> + * @author David Ezzio
>>>>>>>>> + */
>>>>>>>>> +public class TestSerializedFactory
>>>>>>>>> + extends SingleEMFTestCase {
>>>>>>>>> +
>>>>>>>>> + public void setUp() {
>>>>>>>>> + setUp(AllFieldTypes.class);
>>>>>>>>> + }
>>>>>>>>> +
>>>>>>>>> + public void testSerializedEntityManagerFactory() throws
>>>>>>>>> Exception {
>>>>>>>>> + // serialize and deserialize the entity manager
>> factory
>>>>>>>>> + ByteArrayOutputStream baos = new
>> ByteArrayOutputStream();
>>>>>>>>> + ObjectOutputStream oos = new
>> ObjectOutputStream(baos);
>>>>>>>>> + oos.writeObject(emf);
>>>>>>>>> + EntityManagerFactory emf2 =
>>>>>>>>> + (EntityManagerFactory) new ObjectInputStream(
>>>>>>>>> + new
>>>>>>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
>>>>>>>>> +
>>>>>>>>> + // use the deserialized entity manager factory
>>>>>>>>> + assertTrue("The deserialized entity manager factory
>
>>>>>>>>> + is not
>>>>>>>>> open",
>>>>>>>>> + emf2.isOpen());
>>>>>>>>> + EntityManager em = emf2.createEntityManager();
>>>>>>>>> + assertTrue("The newly created entity manager is not
>
>>>>>>>>> + open",
>>>>>>>>> em.isOpen());
>>>>>>>>> +
>>>>>>>>> + // exercise the entity manager produced from the
>>>>>>>>> deserialized EMF
>>>>>>>>> + em.getTransaction().begin();
>>>>>>>>> + em.persist(new AllFieldTypes());
>>>>>>>>> + em.getTransaction().commit();
>>>>>>>>> +
>>>>>>>>> + // close the extra resources
>>>>>>>>> + em.close();
>>>>>>>>> + assertFalse("The entity manager is not closed",
>> em.isOpen());
>>>>>>>>> + emf2.close();
>>>>>>>>> + assertFalse("The entity manager factory is not
>>>>>>>>> + closed",
>>>>>>>>> emf2.isOpen());
>>>>>>>>> + }
>>>>>>>>> +
>>>>>>>>> + public static void main(String[] args) {
>>>>>>>>> + TestRunner.run(TestSerializedFactory.class);
>>>>>>>>> + }
>>>>>>>>> +}
>>>>>>>>> +
>>>>>>>>>
>>>>>>>>> Propchange:
>>>>>>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
>>>>>>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------
>>>>>>>>> --
>>>>>>>>> ----------------
>>>>>>>>>
>>>>>>>>> svn:eol-style = native
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Patrick Linskey
>>>>>>>> 202 669 5907
>>>>>>>>
>>>>>>>
>>>>>>>
>>>
>>>
>>> --
>>> Patrick Linskey
>>> 202 669 5907
>>>
>>
>>
>> --
>> Patrick Linskey
>> 202 669 5907
>>
>> Notice: This email message, together with any attachments, may
>> contain information of BEA Systems, Inc., its subsidiaries and
>> affiliated entities, that may be confidential, proprietary,
>> copyrighted and/or legally privileged, and is intended solely for the
>
>> use of the individual or entity named in this message. If you are not
>> the intended recipient, and have received this message in error,
>> please immediately return this by email and then delete it.
>>
>
> Notice: This email message, together with any attachments, may contain information of BEA Systems, Inc., its subsidiaries and affiliated entities, that may be confidential, proprietary, copyrighted and/or legally privileged, and is intended solely for the use of the individual or entity named in this message. If you are not the intended recipient, and have received this message in error, please immediately return this by email and then delete it.
>
RE: Failures in TeamCity Derby build and svn commit: r616972
Posted by Pinaki Poddar <pp...@bea.com>.
Hi,
Serializablity requirement on EMF is arising out of necessity rather
than by JPA specifcation.
JNDI, too, does not mandate bound objects to be Serializable.
But JNDI providers often place that restriction.
Non-serializable objects can be bound to JNDI by implementing
javax.naming.Referenceable.
Kodo's JDO product derivation based on OpenJPA is following this
Referenceable route.
But I do not see this in openjpa EntityManagerFactory implementation.
-----Original Message-----
From: Kevin Sutter [mailto:kwsutter@gmail.com]
Sent: Friday, February 01, 2008 11:11 AM
To: dev@openjpa.apache.org
Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
On Feb 1, 2008 10:17 AM, Pinaki Poddar <pp...@bea.com> wrote:
> Factories needs to be Serializable as they are often placed in JNDI.
The JNDI lookup needs to work for the Factory definition, but it doesn't
require that the actual Factory be present in JNDI. And, since the JNDI
lookup is meant for the EE (Container) environment, I would guess that
most implementations provide a wrapper object in JNDI for the actual
Factory.
Granted, this wrapper approach is not a requirement. But, I can't find
the requirement in the spec that the EMF has to be Serializable
either... :-)
Kevin
>
> -----Original Message-----
> From: Patrick Linskey [mailto:plinskey@gmail.com]
> Sent: Thursday, January 31, 2008 7:57 PM
> To: dev@openjpa.apache.org
> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
>
> To be honest, I don't really understand why a BrokerFactory needs to
> be serializable, so I think probably it'd be good to make sure that
> there is good reason for that before going too deep into this
> debugging exercise.
>
> -Patrick
>
> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
> > > If the hangup is intolerable, let me know, and I'll back out my
> > > change tomorrow.
> >
> > Personally, I'd rather have the TeamCity build be clean while you
> > investigate. As you noted, the error-reporting from surefire is
> > frustratingly inaccurate, and this really erodes the benefit of
> > continuous integration.
> >
> > > It would be handy if I knew a few things. How to pass in VM debug
> > > args to Maven for its test runs, where to set the OpenJPA logging
> > > when Maven runs the tests, and how to pass in the appropriate
> > > parameters when running the test case outside of Maven. Any
> > > insights on these questions will be appreciated.
> >
> > I usually just add log configuration to the test cases themselves,
> > in the setUp(...) method:
> >
> > setUp(Class.class, Class2.class, ..., "openjpa.Log",
> > "SQL=TRACE");
> >
> > I believe that it is also possible to set this at the command line
> > via
>
> > -Dopenjpa.Log=SQL=TRACE types of clauses.
> >
> > -Patrick
> >
> >
> > On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
> > > Hi,
> > >
> > > Progress has been slow on the test failures apparently introduced
> > > by
>
> > > 616972. I have made progress in reproducing the results.
> > >
> > > The miscounting in TeamCity is gross. Although it lists 33
> > > failing tests complete with a credible stack trace for each, in
> > > fact, there is only one failing test.
> > >
> > > Meanwhile, I believe now, that change 616658 is no longer a
suspect.
>
> > > I get very erratic results here, but they seem to vary more by
> > > whether I run "mvn test" after "mvn clean" or "mvn install" after
> > > "mvn clean". By doing just what TeamCity does, "mvn clean compile
> > > test", I can finally reproduce the results locally that TeamCity
> > > reports in its Compile-Derby build for this change (395, I think).
> > > In spite of this, if I run "mvn
> > > -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
> test passes.
> > >
> > > It would be handy if I knew a few things. How to pass in VM debug
> > > args to Maven for its test runs, where to set the OpenJPA logging
> > > when Maven runs the tests, and how to pass in the appropriate
> > > parameters when running the test case outside of Maven. Any
> > > insights on these questions will be appreciated.
> > >
> > > But at least at this point, I can reproduce locally what TeamCity
> > > reports, and although it happened only once out of one try, I am
> > > optimistic that it is reproducible. I will continue working this
> > > tomorrow.
> > >
> > > If the hangup is intolerable, let me know, and I'll back out my
> > > change tomorrow.
> > >
> > > David Ezzio
> > >
> > >
> > > David Ezzio wrote:
> > > > Hi,
> > > >
> > > > I am convinced that change 616658 (small change to
> > > > DBDictionary.java) is the source of the two intermittent
> > > > failures
> that I am seeing locally.
> > > > One of those intermittent failures is the failure in
> > > > TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
> > > > which is one of the 33 related failures currently seen in
> > > > TeamCity
> Derby builds.
> > > >
> > > > The other intermittent failure is
> > > > TestVersion.testVersionTimestamp
>
> > > > which I reported earlier as occurring locally with change
616658.
> > > >
> > > > When I revert change 616658 locally, both of these issues go
> > > > away even with the changes that I submitted to
> > > > AbstractBrokerFactory in
>
> > > > change 616972.
> > > >
> > > > I have a question: is it possible to run a build in TeamCity
> > > > without submitting the change (or conditionally submitting the
> change) to SVN?
> > > > I'd like to find out if reverting 616658 will fix the issue. If
> > > > so, then I can try to figure out why it's a problem.
> > > >
> > > > Thanks,
> > > >
> > > > David
> > > >
> > > > David Ezzio wrote:
> > > >> No, but I'm taking a look since my changes appear to be the
> cause.
> > > >>
> > > >> Patrick Linskey wrote:
> > > >>> 5. The tests are now failing in the automated build; any idea
> > > >>> why this might be?
> > > >>>
> > > >>> -Patrick
> > > >>>
> > > >>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
> wrote:
> > > >>>> Hi,
> > > >>>>
> > > >>>> A few questions:
> > > >>>>
> > > >>>> 1. Were you seeing this cause problems somewhere? I.e., when
> > > >>>> do
>
> > > >>>> we serialize BrokerFactory instances?
> > > >>>>
> > > >>>> 2. Why not just move the initialization code to the
> > > >>>> declaration
>
> > > >>>> line (instead of the constructor)?
> > > >>>>
> > > >>>> 3. It looks like there are other transactional structures
> > > >>>> that might not be properly handled in serialization /
> > > >>>> deserialization (lifecycle listeners, for example). Should we
> be making more changes there?
> > > >>>>
> > > >>>> -Patrick
> > > >>>>
> > > >>>>
> > > >>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
> > > >>>>> Author: dezzio
> > > >>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
> > > >>>>>
> > > >>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
> > > >>>>> Log:
> > > >>>>> Allow EntityManagerFactory objects to be serialized and
> > > >>>>> deserialized successfully.
> > > >>>>>
> > > >>>>> Added:
> > > >>>>>
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> > > >>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> (with props)
> > > >>>>> Modified:
> > > >>>>>
> > > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
> > > >>>>> a/ kernel/AbstractBrokerFactory.java
> > > >>>>>
> > > >>>>>
> > > >>>>> Modified:
> > > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
> > > >>>>> a/ kernel/AbstractBrokerFactory.java
> > > >>>>>
> > > >>>>> URL:
> > > >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/sr
> > > >>>>> c/
> > > >>>>> main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.ja
> > > >>>>> va ?rev=616972&r1=616971&r2=616972&view=diff
> > > >>>>>
> > > >>>>> ============================================================
> > > >>>>> ==
> > > >>>>> ================
> > > >>>>>
> > > >>>>> ---
> > > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
> > > >>>>> a/ kernel/AbstractBrokerFactory.java
> > > >>>>> (original)
> > > >>>>> +++
> > > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjp
> > > >>>>> a/ kernel/AbstractBrokerFactory.java Wed Jan 30 16:59:02
> > > >>>>> 2008 @@ -147,8 +147,7 @@
> > > >>>>> */
> > > >>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
> config) {
> > > >>>>> _conf = config;
> > > >>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
> > > >>>>> - ConcurrentReferenceHashSet.WEAK);
> > > >>>>> + getPcClassLoaders();
> > > >>>>> }
> > > >>>>>
> > > >>>>> /**
> > > >>>>> @@ -287,13 +286,13 @@
> > > >>>>> if (needsSub(cls))
> > > >>>>> toRedefine.add(cls);
> > > >>>>> }
> > > >>>>> - _pcClassLoaders.add(loader);
> > > >>>>> + getPcClassLoaders().add(loader);
> > > >>>>> _pcClassNames = c;
> > > >>>>> }
> > > >>>>> _persistentTypesLoaded = true;
> > > >>>>> } else {
> > > >>>>> // reload with this loader
> > > >>>>> - if (_pcClassLoaders.add(loader)) {
> > > >>>>> + if (getPcClassLoaders().add(loader)) {
> > > >>>>> for (Iterator itr =
> > > >>>>> _pcClassNames.iterator();
> > > >>>>> itr.hasNext();) {
> > > >>>>> try {
> > > >>>>> Class cls = @@ -818,4 +817,15 @@
> > > >>>>> _transactional.remove (_trans);
> > > >>>>> }
> > > >>>>> }
> > > >>>>> +
> > > >>>>> + /**
> > > >>>>> + * Method insures that deserialized EMF has this
> > > >>>>> + reference
> > > >>>>> re-instantiated
> > > >>>>> + */
> > > >>>>> + private Collection getPcClassLoaders() {
> > > >>>>> + if (_pcClassLoaders == null)
> > > >>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
> > > >>>>> + ConcurrentReferenceHashSet.WEAK);
> > > >>>>> +
> > > >>>>> + return _pcClassLoaders;
> > > >>>>> + }
> > > >>>>> }
> > > >>>>>
> > > >>>>> Added:
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> > > >>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>>
> > > >>>>> URL:
> > > >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persisten
> > > >>>>> ce
> > > >>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/Te
> > > >>>>> st SerializedFactory.java?rev=616972&view=auto
> > > >>>>>
> > > >>>>> ============================================================
> > > >>>>> ==
> > > >>>>> ================
> > > >>>>>
> > > >>>>> ---
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> > > >>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> (added)
> > > >>>>> +++
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> > > >>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> Wed Jan 30 16:59:02 2008
> > > >>>>> @@ -0,0 +1,74 @@
> > > >>>>> +/*
> > > >>>>> + * Licensed to the Apache Software Foundation (ASF) under
> > > >>>>> +one
> > > >>>>> + * or more contributor license agreements. See the NOTICE
> > > >>>>> +file
> > > >>>>> + * distributed with this work for additional information
> > > >>>>> + * regarding copyright ownership. The ASF licenses this
> > > >>>>> +file
> > > >>>>> + * to you 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 org.apache.openjpa.persistence.simple;
> > > >>>>> +
> > > >>>>> +import java.io.*;
> > > >>>>> +
> > > >>>>> +import javax.persistence.EntityManager; import
> > > >>>>> +javax.persistence.EntityManagerFactory;
> > > >>>>> +import javax.persistence.EntityTransaction;
> > > >>>>> +
> > > >>>>> +import junit.textui.TestRunner; import
> > > >>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
> > > >>>>> +import
> > > >>>>> +org.apache.openjpa.persistence.test.SingleEMFTestCase;
> > > >>>>> +
> > > >>>>> +/**
> > > >>>>> + * Tests that a EntityManagerFactory can be used after
> serialization.
> > > >>>>> + *
> > > >>>>> + * @author David Ezzio
> > > >>>>> + */
> > > >>>>> +public class TestSerializedFactory
> > > >>>>> + extends SingleEMFTestCase {
> > > >>>>> +
> > > >>>>> + public void setUp() {
> > > >>>>> + setUp(AllFieldTypes.class);
> > > >>>>> + }
> > > >>>>> +
> > > >>>>> + public void testSerializedEntityManagerFactory() throws
> > > >>>>> Exception {
> > > >>>>> + // serialize and deserialize the entity manager
> factory
> > > >>>>> + ByteArrayOutputStream baos = new
> ByteArrayOutputStream();
> > > >>>>> + ObjectOutputStream oos = new
> ObjectOutputStream(baos);
> > > >>>>> + oos.writeObject(emf);
> > > >>>>> + EntityManagerFactory emf2 =
> > > >>>>> + (EntityManagerFactory) new ObjectInputStream(
> > > >>>>> + new
> > > >>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
> > > >>>>> +
> > > >>>>> + // use the deserialized entity manager factory
> > > >>>>> + assertTrue("The deserialized entity manager factory
> > > >>>>> + is not
> > > >>>>> open",
> > > >>>>> + emf2.isOpen());
> > > >>>>> + EntityManager em = emf2.createEntityManager();
> > > >>>>> + assertTrue("The newly created entity manager is not
> > > >>>>> + open",
> > > >>>>> em.isOpen());
> > > >>>>> +
> > > >>>>> + // exercise the entity manager produced from the
> > > >>>>> deserialized EMF
> > > >>>>> + em.getTransaction().begin();
> > > >>>>> + em.persist(new AllFieldTypes());
> > > >>>>> + em.getTransaction().commit();
> > > >>>>> +
> > > >>>>> + // close the extra resources
> > > >>>>> + em.close();
> > > >>>>> + assertFalse("The entity manager is not closed",
> em.isOpen());
> > > >>>>> + emf2.close();
> > > >>>>> + assertFalse("The entity manager factory is not
> > > >>>>> + closed",
> > > >>>>> emf2.isOpen());
> > > >>>>> + }
> > > >>>>> +
> > > >>>>> + public static void main(String[] args) {
> > > >>>>> + TestRunner.run(TestSerializedFactory.class);
> > > >>>>> + }
> > > >>>>> +}
> > > >>>>> +
> > > >>>>>
> > > >>>>> Propchange:
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apa
> > > >>>>> ch e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>>
> > > >>>>> ------------------------------------------------------------
> > > >>>>> --
> > > >>>>> ----------------
> > > >>>>>
> > > >>>>> svn:eol-style = native
> > > >>>>>
> > > >>>>>
> > > >>>>>
> > > >>>>
> > > >>>>
> > > >>>> --
> > > >>>> Patrick Linskey
> > > >>>> 202 669 5907
> > > >>>>
> > > >>>
> > > >>>
> > > >>>
> > > >>
> > > >
> > >
> >
> >
> >
> > --
> > Patrick Linskey
> > 202 669 5907
> >
>
>
>
> --
> Patrick Linskey
> 202 669 5907
>
> Notice: This email message, together with any attachments, may
> contain information of BEA Systems, Inc., its subsidiaries and
> affiliated entities, that may be confidential, proprietary,
> copyrighted and/or legally privileged, and is intended solely for the
> use of the individual or entity named in this message. If you are not
> the intended recipient, and have received this message in error,
> please immediately return this by email and then delete it.
>
Notice: This email message, together with any attachments, may contain information of BEA Systems, Inc., its subsidiaries and affiliated entities, that may be confidential, proprietary, copyrighted and/or legally privileged, and is intended solely for the use of the individual or entity named in this message. If you are not the intended recipient, and have received this message in error, please immediately return this by email and then delete it.
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by Kevin Sutter <kw...@gmail.com>.
On Feb 1, 2008 10:17 AM, Pinaki Poddar <pp...@bea.com> wrote:
> Factories needs to be Serializable as they are often placed in JNDI.
The JNDI lookup needs to work for the Factory definition, but it doesn't
require that the actual Factory be present in JNDI. And, since the JNDI
lookup is meant for the EE (Container) environment, I would guess that most
implementations provide a wrapper object in JNDI for the actual Factory.
Granted, this wrapper approach is not a requirement. But, I can't find the
requirement in the spec that the EMF has to be Serializable either... :-)
Kevin
>
> -----Original Message-----
> From: Patrick Linskey [mailto:plinskey@gmail.com]
> Sent: Thursday, January 31, 2008 7:57 PM
> To: dev@openjpa.apache.org
> Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
>
> To be honest, I don't really understand why a BrokerFactory needs to be
> serializable, so I think probably it'd be good to make sure that there
> is good reason for that before going too deep into this debugging
> exercise.
>
> -Patrick
>
> On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
> > > If the hangup is intolerable, let me know, and I'll back out my
> > > change tomorrow.
> >
> > Personally, I'd rather have the TeamCity build be clean while you
> > investigate. As you noted, the error-reporting from surefire is
> > frustratingly inaccurate, and this really erodes the benefit of
> > continuous integration.
> >
> > > It would be handy if I knew a few things. How to pass in VM debug
> > > args to Maven for its test runs, where to set the OpenJPA logging
> > > when Maven runs the tests, and how to pass in the appropriate
> > > parameters when running the test case outside of Maven. Any
> > > insights on these questions will be appreciated.
> >
> > I usually just add log configuration to the test cases themselves, in
> > the setUp(...) method:
> >
> > setUp(Class.class, Class2.class, ..., "openjpa.Log", "SQL=TRACE");
> >
> > I believe that it is also possible to set this at the command line via
>
> > -Dopenjpa.Log=SQL=TRACE types of clauses.
> >
> > -Patrick
> >
> >
> > On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
> > > Hi,
> > >
> > > Progress has been slow on the test failures apparently introduced by
>
> > > 616972. I have made progress in reproducing the results.
> > >
> > > The miscounting in TeamCity is gross. Although it lists 33 failing
> > > tests complete with a credible stack trace for each, in fact, there
> > > is only one failing test.
> > >
> > > Meanwhile, I believe now, that change 616658 is no longer a suspect.
>
> > > I get very erratic results here, but they seem to vary more by
> > > whether I run "mvn test" after "mvn clean" or "mvn install" after
> > > "mvn clean". By doing just what TeamCity does, "mvn clean compile
> > > test", I can finally reproduce the results locally that TeamCity
> > > reports in its Compile-Derby build for this change (395, I think).
> > > In spite of this, if I run "mvn
> > > -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
> test passes.
> > >
> > > It would be handy if I knew a few things. How to pass in VM debug
> > > args to Maven for its test runs, where to set the OpenJPA logging
> > > when Maven runs the tests, and how to pass in the appropriate
> > > parameters when running the test case outside of Maven. Any
> > > insights on these questions will be appreciated.
> > >
> > > But at least at this point, I can reproduce locally what TeamCity
> > > reports, and although it happened only once out of one try, I am
> > > optimistic that it is reproducible. I will continue working this
> > > tomorrow.
> > >
> > > If the hangup is intolerable, let me know, and I'll back out my
> > > change tomorrow.
> > >
> > > David Ezzio
> > >
> > >
> > > David Ezzio wrote:
> > > > Hi,
> > > >
> > > > I am convinced that change 616658 (small change to
> > > > DBDictionary.java) is the source of the two intermittent failures
> that I am seeing locally.
> > > > One of those intermittent failures is the failure in
> > > > TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
> > > > which is one of the 33 related failures currently seen in TeamCity
> Derby builds.
> > > >
> > > > The other intermittent failure is TestVersion.testVersionTimestamp
>
> > > > which I reported earlier as occurring locally with change 616658.
> > > >
> > > > When I revert change 616658 locally, both of these issues go away
> > > > even with the changes that I submitted to AbstractBrokerFactory in
>
> > > > change 616972.
> > > >
> > > > I have a question: is it possible to run a build in TeamCity
> > > > without submitting the change (or conditionally submitting the
> change) to SVN?
> > > > I'd like to find out if reverting 616658 will fix the issue. If
> > > > so, then I can try to figure out why it's a problem.
> > > >
> > > > Thanks,
> > > >
> > > > David
> > > >
> > > > David Ezzio wrote:
> > > >> No, but I'm taking a look since my changes appear to be the
> cause.
> > > >>
> > > >> Patrick Linskey wrote:
> > > >>> 5. The tests are now failing in the automated build; any idea
> > > >>> why this might be?
> > > >>>
> > > >>> -Patrick
> > > >>>
> > > >>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
> wrote:
> > > >>>> Hi,
> > > >>>>
> > > >>>> A few questions:
> > > >>>>
> > > >>>> 1. Were you seeing this cause problems somewhere? I.e., when do
>
> > > >>>> we serialize BrokerFactory instances?
> > > >>>>
> > > >>>> 2. Why not just move the initialization code to the declaration
>
> > > >>>> line (instead of the constructor)?
> > > >>>>
> > > >>>> 3. It looks like there are other transactional structures that
> > > >>>> might not be properly handled in serialization /
> > > >>>> deserialization (lifecycle listeners, for example). Should we
> be making more changes there?
> > > >>>>
> > > >>>> -Patrick
> > > >>>>
> > > >>>>
> > > >>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
> > > >>>>> Author: dezzio
> > > >>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
> > > >>>>>
> > > >>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
> > > >>>>> Log:
> > > >>>>> Allow EntityManagerFactory objects to be serialized and
> > > >>>>> deserialized successfully.
> > > >>>>>
> > > >>>>> Added:
> > > >>>>>
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> (with props)
> > > >>>>> Modified:
> > > >>>>>
> > > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/
> > > >>>>> kernel/AbstractBrokerFactory.java
> > > >>>>>
> > > >>>>>
> > > >>>>> Modified:
> > > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/
> > > >>>>> kernel/AbstractBrokerFactory.java
> > > >>>>>
> > > >>>>> URL:
> > > >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/src/
> > > >>>>> main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > > >>>>> ?rev=616972&r1=616971&r2=616972&view=diff
> > > >>>>>
> > > >>>>> ==============================================================
> > > >>>>> ================
> > > >>>>>
> > > >>>>> ---
> > > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/
> > > >>>>> kernel/AbstractBrokerFactory.java
> > > >>>>> (original)
> > > >>>>> +++
> > > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/
> > > >>>>> kernel/AbstractBrokerFactory.java
> > > >>>>> Wed Jan 30 16:59:02 2008
> > > >>>>> @@ -147,8 +147,7 @@
> > > >>>>> */
> > > >>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
> config) {
> > > >>>>> _conf = config;
> > > >>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
> > > >>>>> - ConcurrentReferenceHashSet.WEAK);
> > > >>>>> + getPcClassLoaders();
> > > >>>>> }
> > > >>>>>
> > > >>>>> /**
> > > >>>>> @@ -287,13 +286,13 @@
> > > >>>>> if (needsSub(cls))
> > > >>>>> toRedefine.add(cls);
> > > >>>>> }
> > > >>>>> - _pcClassLoaders.add(loader);
> > > >>>>> + getPcClassLoaders().add(loader);
> > > >>>>> _pcClassNames = c;
> > > >>>>> }
> > > >>>>> _persistentTypesLoaded = true;
> > > >>>>> } else {
> > > >>>>> // reload with this loader
> > > >>>>> - if (_pcClassLoaders.add(loader)) {
> > > >>>>> + if (getPcClassLoaders().add(loader)) {
> > > >>>>> for (Iterator itr = _pcClassNames.iterator();
> > > >>>>> itr.hasNext();) {
> > > >>>>> try {
> > > >>>>> Class cls = @@ -818,4 +817,15 @@
> > > >>>>> _transactional.remove (_trans);
> > > >>>>> }
> > > >>>>> }
> > > >>>>> +
> > > >>>>> + /**
> > > >>>>> + * Method insures that deserialized EMF has this reference
> > > >>>>> re-instantiated
> > > >>>>> + */
> > > >>>>> + private Collection getPcClassLoaders() {
> > > >>>>> + if (_pcClassLoaders == null)
> > > >>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
> > > >>>>> + ConcurrentReferenceHashSet.WEAK);
> > > >>>>> +
> > > >>>>> + return _pcClassLoaders;
> > > >>>>> + }
> > > >>>>> }
> > > >>>>>
> > > >>>>> Added:
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>>
> > > >>>>> URL:
> > > >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persistence
> > > >>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/Test
> > > >>>>> SerializedFactory.java?rev=616972&view=auto
> > > >>>>>
> > > >>>>> ==============================================================
> > > >>>>> ================
> > > >>>>>
> > > >>>>> ---
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> (added)
> > > >>>>> +++
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>> Wed Jan 30 16:59:02 2008
> > > >>>>> @@ -0,0 +1,74 @@
> > > >>>>> +/*
> > > >>>>> + * Licensed to the Apache Software Foundation (ASF) under one
> > > >>>>> + * or more contributor license agreements. See the NOTICE
> > > >>>>> +file
> > > >>>>> + * distributed with this work for additional information
> > > >>>>> + * regarding copyright ownership. The ASF licenses this file
> > > >>>>> + * to you 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 org.apache.openjpa.persistence.simple;
> > > >>>>> +
> > > >>>>> +import java.io.*;
> > > >>>>> +
> > > >>>>> +import javax.persistence.EntityManager; import
> > > >>>>> +javax.persistence.EntityManagerFactory;
> > > >>>>> +import javax.persistence.EntityTransaction;
> > > >>>>> +
> > > >>>>> +import junit.textui.TestRunner; import
> > > >>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
> > > >>>>> +import org.apache.openjpa.persistence.test.SingleEMFTestCase;
> > > >>>>> +
> > > >>>>> +/**
> > > >>>>> + * Tests that a EntityManagerFactory can be used after
> serialization.
> > > >>>>> + *
> > > >>>>> + * @author David Ezzio
> > > >>>>> + */
> > > >>>>> +public class TestSerializedFactory
> > > >>>>> + extends SingleEMFTestCase {
> > > >>>>> +
> > > >>>>> + public void setUp() {
> > > >>>>> + setUp(AllFieldTypes.class);
> > > >>>>> + }
> > > >>>>> +
> > > >>>>> + public void testSerializedEntityManagerFactory() throws
> > > >>>>> Exception {
> > > >>>>> + // serialize and deserialize the entity manager
> factory
> > > >>>>> + ByteArrayOutputStream baos = new
> ByteArrayOutputStream();
> > > >>>>> + ObjectOutputStream oos = new
> ObjectOutputStream(baos);
> > > >>>>> + oos.writeObject(emf);
> > > >>>>> + EntityManagerFactory emf2 =
> > > >>>>> + (EntityManagerFactory) new ObjectInputStream(
> > > >>>>> + new
> > > >>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
> > > >>>>> +
> > > >>>>> + // use the deserialized entity manager factory
> > > >>>>> + assertTrue("The deserialized entity manager factory
> > > >>>>> + is not
> > > >>>>> open",
> > > >>>>> + emf2.isOpen());
> > > >>>>> + EntityManager em = emf2.createEntityManager();
> > > >>>>> + assertTrue("The newly created entity manager is not
> > > >>>>> + open",
> > > >>>>> em.isOpen());
> > > >>>>> +
> > > >>>>> + // exercise the entity manager produced from the
> > > >>>>> deserialized EMF
> > > >>>>> + em.getTransaction().begin();
> > > >>>>> + em.persist(new AllFieldTypes());
> > > >>>>> + em.getTransaction().commit();
> > > >>>>> +
> > > >>>>> + // close the extra resources
> > > >>>>> + em.close();
> > > >>>>> + assertFalse("The entity manager is not closed",
> em.isOpen());
> > > >>>>> + emf2.close();
> > > >>>>> + assertFalse("The entity manager factory is not
> > > >>>>> + closed",
> > > >>>>> emf2.isOpen());
> > > >>>>> + }
> > > >>>>> +
> > > >>>>> + public static void main(String[] args) {
> > > >>>>> + TestRunner.run(TestSerializedFactory.class);
> > > >>>>> + }
> > > >>>>> +}
> > > >>>>> +
> > > >>>>>
> > > >>>>> Propchange:
> > > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > > >>>>>
> > > >>>>> --------------------------------------------------------------
> > > >>>>> ----------------
> > > >>>>>
> > > >>>>> svn:eol-style = native
> > > >>>>>
> > > >>>>>
> > > >>>>>
> > > >>>>
> > > >>>>
> > > >>>> --
> > > >>>> Patrick Linskey
> > > >>>> 202 669 5907
> > > >>>>
> > > >>>
> > > >>>
> > > >>>
> > > >>
> > > >
> > >
> >
> >
> >
> > --
> > Patrick Linskey
> > 202 669 5907
> >
>
>
>
> --
> Patrick Linskey
> 202 669 5907
>
> Notice: This email message, together with any attachments, may contain
> information of BEA Systems, Inc., its subsidiaries and affiliated
> entities, that may be confidential, proprietary, copyrighted and/or
> legally privileged, and is intended solely for the use of the individual or
> entity named in this message. If you are not the intended recipient, and
> have received this message in error, please immediately return this by email
> and then delete it.
>
RE: Failures in TeamCity Derby build and svn commit: r616972
Posted by Pinaki Poddar <pp...@bea.com>.
Factories needs to be Serializable as they are often placed in JNDI.
-----Original Message-----
From: Patrick Linskey [mailto:plinskey@gmail.com]
Sent: Thursday, January 31, 2008 7:57 PM
To: dev@openjpa.apache.org
Subject: Re: Failures in TeamCity Derby build and svn commit: r616972
To be honest, I don't really understand why a BrokerFactory needs to be
serializable, so I think probably it'd be good to make sure that there
is good reason for that before going too deep into this debugging
exercise.
-Patrick
On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
> > If the hangup is intolerable, let me know, and I'll back out my
> > change tomorrow.
>
> Personally, I'd rather have the TeamCity build be clean while you
> investigate. As you noted, the error-reporting from surefire is
> frustratingly inaccurate, and this really erodes the benefit of
> continuous integration.
>
> > It would be handy if I knew a few things. How to pass in VM debug
> > args to Maven for its test runs, where to set the OpenJPA logging
> > when Maven runs the tests, and how to pass in the appropriate
> > parameters when running the test case outside of Maven. Any
> > insights on these questions will be appreciated.
>
> I usually just add log configuration to the test cases themselves, in
> the setUp(...) method:
>
> setUp(Class.class, Class2.class, ..., "openjpa.Log", "SQL=TRACE");
>
> I believe that it is also possible to set this at the command line via
> -Dopenjpa.Log=SQL=TRACE types of clauses.
>
> -Patrick
>
>
> On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
> > Hi,
> >
> > Progress has been slow on the test failures apparently introduced by
> > 616972. I have made progress in reproducing the results.
> >
> > The miscounting in TeamCity is gross. Although it lists 33 failing
> > tests complete with a credible stack trace for each, in fact, there
> > is only one failing test.
> >
> > Meanwhile, I believe now, that change 616658 is no longer a suspect.
> > I get very erratic results here, but they seem to vary more by
> > whether I run "mvn test" after "mvn clean" or "mvn install" after
> > "mvn clean". By doing just what TeamCity does, "mvn clean compile
> > test", I can finally reproduce the results locally that TeamCity
> > reports in its Compile-Derby build for this change (395, I think).
> > In spite of this, if I run "mvn
> > -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the
test passes.
> >
> > It would be handy if I knew a few things. How to pass in VM debug
> > args to Maven for its test runs, where to set the OpenJPA logging
> > when Maven runs the tests, and how to pass in the appropriate
> > parameters when running the test case outside of Maven. Any
> > insights on these questions will be appreciated.
> >
> > But at least at this point, I can reproduce locally what TeamCity
> > reports, and although it happened only once out of one try, I am
> > optimistic that it is reproducible. I will continue working this
> > tomorrow.
> >
> > If the hangup is intolerable, let me know, and I'll back out my
> > change tomorrow.
> >
> > David Ezzio
> >
> >
> > David Ezzio wrote:
> > > Hi,
> > >
> > > I am convinced that change 616658 (small change to
> > > DBDictionary.java) is the source of the two intermittent failures
that I am seeing locally.
> > > One of those intermittent failures is the failure in
> > > TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans
> > > which is one of the 33 related failures currently seen in TeamCity
Derby builds.
> > >
> > > The other intermittent failure is TestVersion.testVersionTimestamp
> > > which I reported earlier as occurring locally with change 616658.
> > >
> > > When I revert change 616658 locally, both of these issues go away
> > > even with the changes that I submitted to AbstractBrokerFactory in
> > > change 616972.
> > >
> > > I have a question: is it possible to run a build in TeamCity
> > > without submitting the change (or conditionally submitting the
change) to SVN?
> > > I'd like to find out if reverting 616658 will fix the issue. If
> > > so, then I can try to figure out why it's a problem.
> > >
> > > Thanks,
> > >
> > > David
> > >
> > > David Ezzio wrote:
> > >> No, but I'm taking a look since my changes appear to be the
cause.
> > >>
> > >> Patrick Linskey wrote:
> > >>> 5. The tests are now failing in the automated build; any idea
> > >>> why this might be?
> > >>>
> > >>> -Patrick
> > >>>
> > >>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com>
wrote:
> > >>>> Hi,
> > >>>>
> > >>>> A few questions:
> > >>>>
> > >>>> 1. Were you seeing this cause problems somewhere? I.e., when do
> > >>>> we serialize BrokerFactory instances?
> > >>>>
> > >>>> 2. Why not just move the initialization code to the declaration
> > >>>> line (instead of the constructor)?
> > >>>>
> > >>>> 3. It looks like there are other transactional structures that
> > >>>> might not be properly handled in serialization /
> > >>>> deserialization (lifecycle listeners, for example). Should we
be making more changes there?
> > >>>>
> > >>>> -Patrick
> > >>>>
> > >>>>
> > >>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
> > >>>>> Author: dezzio
> > >>>>> Date: Wed Jan 30 16:59:02 2008 New Revision: 616972
> > >>>>>
> > >>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
> > >>>>> Log:
> > >>>>> Allow EntityManagerFactory objects to be serialized and
> > >>>>> deserialized successfully.
> > >>>>>
> > >>>>> Added:
> > >>>>>
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>> (with props)
> > >>>>> Modified:
> > >>>>>
> > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/
> > >>>>> kernel/AbstractBrokerFactory.java
> > >>>>>
> > >>>>>
> > >>>>> Modified:
> > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/
> > >>>>> kernel/AbstractBrokerFactory.java
> > >>>>>
> > >>>>> URL:
> > >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/src/
> > >>>>> main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > >>>>> ?rev=616972&r1=616971&r2=616972&view=diff
> > >>>>>
> > >>>>> ==============================================================
> > >>>>> ================
> > >>>>>
> > >>>>> ---
> > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/
> > >>>>> kernel/AbstractBrokerFactory.java
> > >>>>> (original)
> > >>>>> +++
> > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/
> > >>>>> kernel/AbstractBrokerFactory.java
> > >>>>> Wed Jan 30 16:59:02 2008
> > >>>>> @@ -147,8 +147,7 @@
> > >>>>> */
> > >>>>> protected AbstractBrokerFactory(OpenJPAConfiguration
config) {
> > >>>>> _conf = config;
> > >>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
> > >>>>> - ConcurrentReferenceHashSet.WEAK);
> > >>>>> + getPcClassLoaders();
> > >>>>> }
> > >>>>>
> > >>>>> /**
> > >>>>> @@ -287,13 +286,13 @@
> > >>>>> if (needsSub(cls))
> > >>>>> toRedefine.add(cls);
> > >>>>> }
> > >>>>> - _pcClassLoaders.add(loader);
> > >>>>> + getPcClassLoaders().add(loader);
> > >>>>> _pcClassNames = c;
> > >>>>> }
> > >>>>> _persistentTypesLoaded = true;
> > >>>>> } else {
> > >>>>> // reload with this loader
> > >>>>> - if (_pcClassLoaders.add(loader)) {
> > >>>>> + if (getPcClassLoaders().add(loader)) {
> > >>>>> for (Iterator itr = _pcClassNames.iterator();
> > >>>>> itr.hasNext();) {
> > >>>>> try {
> > >>>>> Class cls = @@ -818,4 +817,15 @@
> > >>>>> _transactional.remove (_trans);
> > >>>>> }
> > >>>>> }
> > >>>>> +
> > >>>>> + /**
> > >>>>> + * Method insures that deserialized EMF has this reference
> > >>>>> re-instantiated
> > >>>>> + */
> > >>>>> + private Collection getPcClassLoaders() {
> > >>>>> + if (_pcClassLoaders == null)
> > >>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
> > >>>>> + ConcurrentReferenceHashSet.WEAK);
> > >>>>> +
> > >>>>> + return _pcClassLoaders;
> > >>>>> + }
> > >>>>> }
> > >>>>>
> > >>>>> Added:
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>>
> > >>>>> URL:
> > >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persistence
> > >>>>> -jdbc/src/test/java/org/apache/openjpa/persistence/simple/Test
> > >>>>> SerializedFactory.java?rev=616972&view=auto
> > >>>>>
> > >>>>> ==============================================================
> > >>>>> ================
> > >>>>>
> > >>>>> ---
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>> (added)
> > >>>>> +++
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>> Wed Jan 30 16:59:02 2008
> > >>>>> @@ -0,0 +1,74 @@
> > >>>>> +/*
> > >>>>> + * Licensed to the Apache Software Foundation (ASF) under one
> > >>>>> + * or more contributor license agreements. See the NOTICE
> > >>>>> +file
> > >>>>> + * distributed with this work for additional information
> > >>>>> + * regarding copyright ownership. The ASF licenses this file
> > >>>>> + * to you 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 org.apache.openjpa.persistence.simple;
> > >>>>> +
> > >>>>> +import java.io.*;
> > >>>>> +
> > >>>>> +import javax.persistence.EntityManager; import
> > >>>>> +javax.persistence.EntityManagerFactory;
> > >>>>> +import javax.persistence.EntityTransaction;
> > >>>>> +
> > >>>>> +import junit.textui.TestRunner; import
> > >>>>> +org.apache.openjpa.persistence.OpenJPAEntityManager;
> > >>>>> +import org.apache.openjpa.persistence.test.SingleEMFTestCase;
> > >>>>> +
> > >>>>> +/**
> > >>>>> + * Tests that a EntityManagerFactory can be used after
serialization.
> > >>>>> + *
> > >>>>> + * @author David Ezzio
> > >>>>> + */
> > >>>>> +public class TestSerializedFactory
> > >>>>> + extends SingleEMFTestCase {
> > >>>>> +
> > >>>>> + public void setUp() {
> > >>>>> + setUp(AllFieldTypes.class);
> > >>>>> + }
> > >>>>> +
> > >>>>> + public void testSerializedEntityManagerFactory() throws
> > >>>>> Exception {
> > >>>>> + // serialize and deserialize the entity manager
factory
> > >>>>> + ByteArrayOutputStream baos = new
ByteArrayOutputStream();
> > >>>>> + ObjectOutputStream oos = new
ObjectOutputStream(baos);
> > >>>>> + oos.writeObject(emf);
> > >>>>> + EntityManagerFactory emf2 =
> > >>>>> + (EntityManagerFactory) new ObjectInputStream(
> > >>>>> + new
> > >>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
> > >>>>> +
> > >>>>> + // use the deserialized entity manager factory
> > >>>>> + assertTrue("The deserialized entity manager factory
> > >>>>> + is not
> > >>>>> open",
> > >>>>> + emf2.isOpen());
> > >>>>> + EntityManager em = emf2.createEntityManager();
> > >>>>> + assertTrue("The newly created entity manager is not
> > >>>>> + open",
> > >>>>> em.isOpen());
> > >>>>> +
> > >>>>> + // exercise the entity manager produced from the
> > >>>>> deserialized EMF
> > >>>>> + em.getTransaction().begin();
> > >>>>> + em.persist(new AllFieldTypes());
> > >>>>> + em.getTransaction().commit();
> > >>>>> +
> > >>>>> + // close the extra resources
> > >>>>> + em.close();
> > >>>>> + assertFalse("The entity manager is not closed",
em.isOpen());
> > >>>>> + emf2.close();
> > >>>>> + assertFalse("The entity manager factory is not
> > >>>>> + closed",
> > >>>>> emf2.isOpen());
> > >>>>> + }
> > >>>>> +
> > >>>>> + public static void main(String[] args) {
> > >>>>> + TestRunner.run(TestSerializedFactory.class);
> > >>>>> + }
> > >>>>> +}
> > >>>>> +
> > >>>>>
> > >>>>> Propchange:
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apach
> > >>>>> e/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>>
> > >>>>> --------------------------------------------------------------
> > >>>>> ----------------
> > >>>>>
> > >>>>> svn:eol-style = native
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>>
> > >>>>
> > >>>> --
> > >>>> Patrick Linskey
> > >>>> 202 669 5907
> > >>>>
> > >>>
> > >>>
> > >>>
> > >>
> > >
> >
>
>
>
> --
> Patrick Linskey
> 202 669 5907
>
--
Patrick Linskey
202 669 5907
Notice: This email message, together with any attachments, may contain information of BEA Systems, Inc., its subsidiaries and affiliated entities, that may be confidential, proprietary, copyrighted and/or legally privileged, and is intended solely for the use of the individual or entity named in this message. If you are not the intended recipient, and have received this message in error, please immediately return this by email and then delete it.
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by Patrick Linskey <pl...@gmail.com>.
To be honest, I don't really understand why a BrokerFactory needs to
be serializable, so I think probably it'd be good to make sure that
there is good reason for that before going too deep into this
debugging exercise.
-Patrick
On Jan 31, 2008 5:52 PM, Patrick Linskey <pl...@gmail.com> wrote:
> > If the hangup is intolerable, let me know, and I'll back out my change
> > tomorrow.
>
> Personally, I'd rather have the TeamCity build be clean while you
> investigate. As you noted, the error-reporting from surefire is
> frustratingly inaccurate, and this really erodes the benefit of
> continuous integration.
>
> > It would be handy if I knew a few things. How to pass in VM debug args
> > to Maven for its test runs, where to set the OpenJPA logging when Maven
> > runs the tests, and how to pass in the appropriate parameters when
> > running the test case outside of Maven. Any insights on these questions
> > will be appreciated.
>
> I usually just add log configuration to the test cases themselves, in
> the setUp(...) method:
>
> setUp(Class.class, Class2.class, ..., "openjpa.Log", "SQL=TRACE");
>
> I believe that it is also possible to set this at the command line via
> -Dopenjpa.Log=SQL=TRACE types of clauses.
>
> -Patrick
>
>
> On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
> > Hi,
> >
> > Progress has been slow on the test failures apparently introduced by
> > 616972. I have made progress in reproducing the results.
> >
> > The miscounting in TeamCity is gross. Although it lists 33 failing
> > tests complete with a credible stack trace for each, in fact, there is
> > only one failing test.
> >
> > Meanwhile, I believe now, that change 616658 is no longer a suspect. I
> > get very erratic results here, but they seem to vary more by whether I
> > run "mvn test" after "mvn clean" or "mvn install" after "mvn clean". By
> > doing just what TeamCity does, "mvn clean compile test", I can finally
> > reproduce the results locally that TeamCity reports in its Compile-Derby
> > build for this change (395, I think). In spite of this, if I run "mvn
> > -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the test
> > passes.
> >
> > It would be handy if I knew a few things. How to pass in VM debug args
> > to Maven for its test runs, where to set the OpenJPA logging when Maven
> > runs the tests, and how to pass in the appropriate parameters when
> > running the test case outside of Maven. Any insights on these questions
> > will be appreciated.
> >
> > But at least at this point, I can reproduce locally what TeamCity
> > reports, and although it happened only once out of one try, I am
> > optimistic that it is reproducible. I will continue working this
> > tomorrow.
> >
> > If the hangup is intolerable, let me know, and I'll back out my change
> > tomorrow.
> >
> > David Ezzio
> >
> >
> > David Ezzio wrote:
> > > Hi,
> > >
> > > I am convinced that change 616658 (small change to DBDictionary.java) is
> > > the source of the two intermittent failures that I am seeing locally.
> > > One of those intermittent failures is the failure in
> > > TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans which is
> > > one of the 33 related failures currently seen in TeamCity Derby builds.
> > >
> > > The other intermittent failure is TestVersion.testVersionTimestamp which
> > > I reported earlier as occurring locally with change 616658.
> > >
> > > When I revert change 616658 locally, both of these issues go away even
> > > with the changes that I submitted to AbstractBrokerFactory in change
> > > 616972.
> > >
> > > I have a question: is it possible to run a build in TeamCity without
> > > submitting the change (or conditionally submitting the change) to SVN?
> > > I'd like to find out if reverting 616658 will fix the issue. If so,
> > > then I can try to figure out why it's a problem.
> > >
> > > Thanks,
> > >
> > > David
> > >
> > > David Ezzio wrote:
> > >> No, but I'm taking a look since my changes appear to be the cause.
> > >>
> > >> Patrick Linskey wrote:
> > >>> 5. The tests are now failing in the automated build; any idea why
> > >>> this might be?
> > >>>
> > >>> -Patrick
> > >>>
> > >>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com> wrote:
> > >>>> Hi,
> > >>>>
> > >>>> A few questions:
> > >>>>
> > >>>> 1. Were you seeing this cause problems somewhere? I.e., when do we
> > >>>> serialize BrokerFactory instances?
> > >>>>
> > >>>> 2. Why not just move the initialization code to the declaration line
> > >>>> (instead of the constructor)?
> > >>>>
> > >>>> 3. It looks like there are other transactional structures that might
> > >>>> not be properly handled in serialization / deserialization (lifecycle
> > >>>> listeners, for example). Should we be making more changes there?
> > >>>>
> > >>>> -Patrick
> > >>>>
> > >>>>
> > >>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
> > >>>>> Author: dezzio
> > >>>>> Date: Wed Jan 30 16:59:02 2008
> > >>>>> New Revision: 616972
> > >>>>>
> > >>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
> > >>>>> Log:
> > >>>>> Allow EntityManagerFactory objects to be serialized and
> > >>>>> deserialized successfully.
> > >>>>>
> > >>>>> Added:
> > >>>>>
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>> (with props)
> > >>>>> Modified:
> > >>>>>
> > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > >>>>>
> > >>>>>
> > >>>>> Modified:
> > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > >>>>>
> > >>>>> URL:
> > >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java?rev=616972&r1=616971&r2=616972&view=diff
> > >>>>>
> > >>>>> ==============================================================================
> > >>>>>
> > >>>>> ---
> > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > >>>>> (original)
> > >>>>> +++
> > >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> > >>>>> Wed Jan 30 16:59:02 2008
> > >>>>> @@ -147,8 +147,7 @@
> > >>>>> */
> > >>>>> protected AbstractBrokerFactory(OpenJPAConfiguration config) {
> > >>>>> _conf = config;
> > >>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
> > >>>>> - ConcurrentReferenceHashSet.WEAK);
> > >>>>> + getPcClassLoaders();
> > >>>>> }
> > >>>>>
> > >>>>> /**
> > >>>>> @@ -287,13 +286,13 @@
> > >>>>> if (needsSub(cls))
> > >>>>> toRedefine.add(cls);
> > >>>>> }
> > >>>>> - _pcClassLoaders.add(loader);
> > >>>>> + getPcClassLoaders().add(loader);
> > >>>>> _pcClassNames = c;
> > >>>>> }
> > >>>>> _persistentTypesLoaded = true;
> > >>>>> } else {
> > >>>>> // reload with this loader
> > >>>>> - if (_pcClassLoaders.add(loader)) {
> > >>>>> + if (getPcClassLoaders().add(loader)) {
> > >>>>> for (Iterator itr = _pcClassNames.iterator();
> > >>>>> itr.hasNext();) {
> > >>>>> try {
> > >>>>> Class cls =
> > >>>>> @@ -818,4 +817,15 @@
> > >>>>> _transactional.remove (_trans);
> > >>>>> }
> > >>>>> }
> > >>>>> +
> > >>>>> + /**
> > >>>>> + * Method insures that deserialized EMF has this reference
> > >>>>> re-instantiated
> > >>>>> + */
> > >>>>> + private Collection getPcClassLoaders() {
> > >>>>> + if (_pcClassLoaders == null)
> > >>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
> > >>>>> + ConcurrentReferenceHashSet.WEAK);
> > >>>>> +
> > >>>>> + return _pcClassLoaders;
> > >>>>> + }
> > >>>>> }
> > >>>>>
> > >>>>> Added:
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>>
> > >>>>> URL:
> > >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java?rev=616972&view=auto
> > >>>>>
> > >>>>> ==============================================================================
> > >>>>>
> > >>>>> ---
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>> (added)
> > >>>>> +++
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>> Wed Jan 30 16:59:02 2008
> > >>>>> @@ -0,0 +1,74 @@
> > >>>>> +/*
> > >>>>> + * Licensed to the Apache Software Foundation (ASF) under one
> > >>>>> + * or more contributor license agreements. See the NOTICE file
> > >>>>> + * distributed with this work for additional information
> > >>>>> + * regarding copyright ownership. The ASF licenses this file
> > >>>>> + * to you 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 org.apache.openjpa.persistence.simple;
> > >>>>> +
> > >>>>> +import java.io.*;
> > >>>>> +
> > >>>>> +import javax.persistence.EntityManager;
> > >>>>> +import javax.persistence.EntityManagerFactory;
> > >>>>> +import javax.persistence.EntityTransaction;
> > >>>>> +
> > >>>>> +import junit.textui.TestRunner;
> > >>>>> +import org.apache.openjpa.persistence.OpenJPAEntityManager;
> > >>>>> +import org.apache.openjpa.persistence.test.SingleEMFTestCase;
> > >>>>> +
> > >>>>> +/**
> > >>>>> + * Tests that a EntityManagerFactory can be used after serialization.
> > >>>>> + *
> > >>>>> + * @author David Ezzio
> > >>>>> + */
> > >>>>> +public class TestSerializedFactory
> > >>>>> + extends SingleEMFTestCase {
> > >>>>> +
> > >>>>> + public void setUp() {
> > >>>>> + setUp(AllFieldTypes.class);
> > >>>>> + }
> > >>>>> +
> > >>>>> + public void testSerializedEntityManagerFactory() throws
> > >>>>> Exception {
> > >>>>> + // serialize and deserialize the entity manager factory
> > >>>>> + ByteArrayOutputStream baos = new ByteArrayOutputStream();
> > >>>>> + ObjectOutputStream oos = new ObjectOutputStream(baos);
> > >>>>> + oos.writeObject(emf);
> > >>>>> + EntityManagerFactory emf2 =
> > >>>>> + (EntityManagerFactory) new ObjectInputStream(
> > >>>>> + new
> > >>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
> > >>>>> +
> > >>>>> + // use the deserialized entity manager factory
> > >>>>> + assertTrue("The deserialized entity manager factory is not
> > >>>>> open",
> > >>>>> + emf2.isOpen());
> > >>>>> + EntityManager em = emf2.createEntityManager();
> > >>>>> + assertTrue("The newly created entity manager is not open",
> > >>>>> em.isOpen());
> > >>>>> +
> > >>>>> + // exercise the entity manager produced from the
> > >>>>> deserialized EMF
> > >>>>> + em.getTransaction().begin();
> > >>>>> + em.persist(new AllFieldTypes());
> > >>>>> + em.getTransaction().commit();
> > >>>>> +
> > >>>>> + // close the extra resources
> > >>>>> + em.close();
> > >>>>> + assertFalse("The entity manager is not closed", em.isOpen());
> > >>>>> + emf2.close();
> > >>>>> + assertFalse("The entity manager factory is not closed",
> > >>>>> emf2.isOpen());
> > >>>>> + }
> > >>>>> +
> > >>>>> + public static void main(String[] args) {
> > >>>>> + TestRunner.run(TestSerializedFactory.class);
> > >>>>> + }
> > >>>>> +}
> > >>>>> +
> > >>>>>
> > >>>>> Propchange:
> > >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> > >>>>>
> > >>>>> ------------------------------------------------------------------------------
> > >>>>>
> > >>>>> svn:eol-style = native
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>>
> > >>>>
> > >>>> --
> > >>>> Patrick Linskey
> > >>>> 202 669 5907
> > >>>>
> > >>>
> > >>>
> > >>>
> > >>
> > >
> >
>
>
>
> --
> Patrick Linskey
> 202 669 5907
>
--
Patrick Linskey
202 669 5907
Re: Failures in TeamCity Derby build and svn commit: r616972
Posted by Patrick Linskey <pl...@gmail.com>.
> If the hangup is intolerable, let me know, and I'll back out my change
> tomorrow.
Personally, I'd rather have the TeamCity build be clean while you
investigate. As you noted, the error-reporting from surefire is
frustratingly inaccurate, and this really erodes the benefit of
continuous integration.
> It would be handy if I knew a few things. How to pass in VM debug args
> to Maven for its test runs, where to set the OpenJPA logging when Maven
> runs the tests, and how to pass in the appropriate parameters when
> running the test case outside of Maven. Any insights on these questions
> will be appreciated.
I usually just add log configuration to the test cases themselves, in
the setUp(...) method:
setUp(Class.class, Class2.class, ..., "openjpa.Log", "SQL=TRACE");
I believe that it is also possible to set this at the command line via
-Dopenjpa.Log=SQL=TRACE types of clauses.
-Patrick
On Jan 31, 2008 4:08 PM, David Ezzio <de...@apache.org> wrote:
> Hi,
>
> Progress has been slow on the test failures apparently introduced by
> 616972. I have made progress in reproducing the results.
>
> The miscounting in TeamCity is gross. Although it lists 33 failing
> tests complete with a credible stack trace for each, in fact, there is
> only one failing test.
>
> Meanwhile, I believe now, that change 616658 is no longer a suspect. I
> get very erratic results here, but they seem to vary more by whether I
> run "mvn test" after "mvn clean" or "mvn install" after "mvn clean". By
> doing just what TeamCity does, "mvn clean compile test", I can finally
> reproduce the results locally that TeamCity reports in its Compile-Derby
> build for this change (395, I think). In spite of this, if I run "mvn
> -Dtest=TestCaseInsensitiveKeywordsInJPQL clean compile test" the test
> passes.
>
> It would be handy if I knew a few things. How to pass in VM debug args
> to Maven for its test runs, where to set the OpenJPA logging when Maven
> runs the tests, and how to pass in the appropriate parameters when
> running the test case outside of Maven. Any insights on these questions
> will be appreciated.
>
> But at least at this point, I can reproduce locally what TeamCity
> reports, and although it happened only once out of one try, I am
> optimistic that it is reproducible. I will continue working this
> tomorrow.
>
> If the hangup is intolerable, let me know, and I'll back out my change
> tomorrow.
>
> David Ezzio
>
>
> David Ezzio wrote:
> > Hi,
> >
> > I am convinced that change 616658 (small change to DBDictionary.java) is
> > the source of the two intermittent failures that I am seeing locally.
> > One of those intermittent failures is the failure in
> > TestCaseInsensitiveKeywordsInJPQL.testCaseInsensitiveBooleans which is
> > one of the 33 related failures currently seen in TeamCity Derby builds.
> >
> > The other intermittent failure is TestVersion.testVersionTimestamp which
> > I reported earlier as occurring locally with change 616658.
> >
> > When I revert change 616658 locally, both of these issues go away even
> > with the changes that I submitted to AbstractBrokerFactory in change
> > 616972.
> >
> > I have a question: is it possible to run a build in TeamCity without
> > submitting the change (or conditionally submitting the change) to SVN?
> > I'd like to find out if reverting 616658 will fix the issue. If so,
> > then I can try to figure out why it's a problem.
> >
> > Thanks,
> >
> > David
> >
> > David Ezzio wrote:
> >> No, but I'm taking a look since my changes appear to be the cause.
> >>
> >> Patrick Linskey wrote:
> >>> 5. The tests are now failing in the automated build; any idea why
> >>> this might be?
> >>>
> >>> -Patrick
> >>>
> >>> On Jan 30, 2008 6:11 PM, Patrick Linskey <pl...@gmail.com> wrote:
> >>>> Hi,
> >>>>
> >>>> A few questions:
> >>>>
> >>>> 1. Were you seeing this cause problems somewhere? I.e., when do we
> >>>> serialize BrokerFactory instances?
> >>>>
> >>>> 2. Why not just move the initialization code to the declaration line
> >>>> (instead of the constructor)?
> >>>>
> >>>> 3. It looks like there are other transactional structures that might
> >>>> not be properly handled in serialization / deserialization (lifecycle
> >>>> listeners, for example). Should we be making more changes there?
> >>>>
> >>>> -Patrick
> >>>>
> >>>>
> >>>> On Jan 30, 2008 4:59 PM, <de...@apache.org> wrote:
> >>>>> Author: dezzio
> >>>>> Date: Wed Jan 30 16:59:02 2008
> >>>>> New Revision: 616972
> >>>>>
> >>>>> URL: http://svn.apache.org/viewvc?rev=616972&view=rev
> >>>>> Log:
> >>>>> Allow EntityManagerFactory objects to be serialized and
> >>>>> deserialized successfully.
> >>>>>
> >>>>> Added:
> >>>>>
> >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>> (with props)
> >>>>> Modified:
> >>>>>
> >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> >>>>>
> >>>>>
> >>>>> Modified:
> >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> >>>>>
> >>>>> URL:
> >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java?rev=616972&r1=616971&r2=616972&view=diff
> >>>>>
> >>>>> ==============================================================================
> >>>>>
> >>>>> ---
> >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> >>>>> (original)
> >>>>> +++
> >>>>> openjpa/trunk/openjpa-kernel/src/main/java/org/apache/openjpa/kernel/AbstractBrokerFactory.java
> >>>>> Wed Jan 30 16:59:02 2008
> >>>>> @@ -147,8 +147,7 @@
> >>>>> */
> >>>>> protected AbstractBrokerFactory(OpenJPAConfiguration config) {
> >>>>> _conf = config;
> >>>>> - _pcClassLoaders = new ConcurrentReferenceHashSet(
> >>>>> - ConcurrentReferenceHashSet.WEAK);
> >>>>> + getPcClassLoaders();
> >>>>> }
> >>>>>
> >>>>> /**
> >>>>> @@ -287,13 +286,13 @@
> >>>>> if (needsSub(cls))
> >>>>> toRedefine.add(cls);
> >>>>> }
> >>>>> - _pcClassLoaders.add(loader);
> >>>>> + getPcClassLoaders().add(loader);
> >>>>> _pcClassNames = c;
> >>>>> }
> >>>>> _persistentTypesLoaded = true;
> >>>>> } else {
> >>>>> // reload with this loader
> >>>>> - if (_pcClassLoaders.add(loader)) {
> >>>>> + if (getPcClassLoaders().add(loader)) {
> >>>>> for (Iterator itr = _pcClassNames.iterator();
> >>>>> itr.hasNext();) {
> >>>>> try {
> >>>>> Class cls =
> >>>>> @@ -818,4 +817,15 @@
> >>>>> _transactional.remove (_trans);
> >>>>> }
> >>>>> }
> >>>>> +
> >>>>> + /**
> >>>>> + * Method insures that deserialized EMF has this reference
> >>>>> re-instantiated
> >>>>> + */
> >>>>> + private Collection getPcClassLoaders() {
> >>>>> + if (_pcClassLoaders == null)
> >>>>> + _pcClassLoaders = new ConcurrentReferenceHashSet(
> >>>>> + ConcurrentReferenceHashSet.WEAK);
> >>>>> +
> >>>>> + return _pcClassLoaders;
> >>>>> + }
> >>>>> }
> >>>>>
> >>>>> Added:
> >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>
> >>>>> URL:
> >>>>> http://svn.apache.org/viewvc/openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java?rev=616972&view=auto
> >>>>>
> >>>>> ==============================================================================
> >>>>>
> >>>>> ---
> >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>> (added)
> >>>>> +++
> >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>> Wed Jan 30 16:59:02 2008
> >>>>> @@ -0,0 +1,74 @@
> >>>>> +/*
> >>>>> + * Licensed to the Apache Software Foundation (ASF) under one
> >>>>> + * or more contributor license agreements. See the NOTICE file
> >>>>> + * distributed with this work for additional information
> >>>>> + * regarding copyright ownership. The ASF licenses this file
> >>>>> + * to you 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 org.apache.openjpa.persistence.simple;
> >>>>> +
> >>>>> +import java.io.*;
> >>>>> +
> >>>>> +import javax.persistence.EntityManager;
> >>>>> +import javax.persistence.EntityManagerFactory;
> >>>>> +import javax.persistence.EntityTransaction;
> >>>>> +
> >>>>> +import junit.textui.TestRunner;
> >>>>> +import org.apache.openjpa.persistence.OpenJPAEntityManager;
> >>>>> +import org.apache.openjpa.persistence.test.SingleEMFTestCase;
> >>>>> +
> >>>>> +/**
> >>>>> + * Tests that a EntityManagerFactory can be used after serialization.
> >>>>> + *
> >>>>> + * @author David Ezzio
> >>>>> + */
> >>>>> +public class TestSerializedFactory
> >>>>> + extends SingleEMFTestCase {
> >>>>> +
> >>>>> + public void setUp() {
> >>>>> + setUp(AllFieldTypes.class);
> >>>>> + }
> >>>>> +
> >>>>> + public void testSerializedEntityManagerFactory() throws
> >>>>> Exception {
> >>>>> + // serialize and deserialize the entity manager factory
> >>>>> + ByteArrayOutputStream baos = new ByteArrayOutputStream();
> >>>>> + ObjectOutputStream oos = new ObjectOutputStream(baos);
> >>>>> + oos.writeObject(emf);
> >>>>> + EntityManagerFactory emf2 =
> >>>>> + (EntityManagerFactory) new ObjectInputStream(
> >>>>> + new
> >>>>> ByteArrayInputStream(baos.toByteArray())).readObject();
> >>>>> +
> >>>>> + // use the deserialized entity manager factory
> >>>>> + assertTrue("The deserialized entity manager factory is not
> >>>>> open",
> >>>>> + emf2.isOpen());
> >>>>> + EntityManager em = emf2.createEntityManager();
> >>>>> + assertTrue("The newly created entity manager is not open",
> >>>>> em.isOpen());
> >>>>> +
> >>>>> + // exercise the entity manager produced from the
> >>>>> deserialized EMF
> >>>>> + em.getTransaction().begin();
> >>>>> + em.persist(new AllFieldTypes());
> >>>>> + em.getTransaction().commit();
> >>>>> +
> >>>>> + // close the extra resources
> >>>>> + em.close();
> >>>>> + assertFalse("The entity manager is not closed", em.isOpen());
> >>>>> + emf2.close();
> >>>>> + assertFalse("The entity manager factory is not closed",
> >>>>> emf2.isOpen());
> >>>>> + }
> >>>>> +
> >>>>> + public static void main(String[] args) {
> >>>>> + TestRunner.run(TestSerializedFactory.class);
> >>>>> + }
> >>>>> +}
> >>>>> +
> >>>>>
> >>>>> Propchange:
> >>>>> openjpa/trunk/openjpa-persistence-jdbc/src/test/java/org/apache/openjpa/persistence/simple/TestSerializedFactory.java
> >>>>>
> >>>>> ------------------------------------------------------------------------------
> >>>>>
> >>>>> svn:eol-style = native
> >>>>>
> >>>>>
> >>>>>
> >>>>
> >>>>
> >>>> --
> >>>> Patrick Linskey
> >>>> 202 669 5907
> >>>>
> >>>
> >>>
> >>>
> >>
> >
>
--
Patrick Linskey
202 669 5907