[hibernate-dev] Regression after upgrade to 4.1.7.Final

[Shawn Clowater]

Spent some time on this.  It's not my change, it's the change that
allows for collections to be loaded outside of a session.

In AbstractPersistentCollection
				boolean isExtraLazy =
withTemporarySessionIfNeeded(
						new
LazyInitializationWork<Boolean>() {
							@Override
							public Boolean
doWork() {
	
CollectionEntry entry =
session.getPersistenceContext().getCollectionEntry(
AbstractPersistentCollection.this );
	
CollectionPersister persister = entry.getLoadedPersister();
								if (
persister.isExtraLazy() ) {
	
if ( hasQueuedOperations() ) {
	
session.flush();
	
}
	
cachedSize = persister.getSize( entry.getLoadedKey(), session );
	
return true;
								}
								else {
	
read();
								}
								return
false;
							}
						}
				);

In this case the entry variable ends up being null and then it blows a
NPE out of here where it used to blow a LazyInitializationException.
Search is checking for that exception to stop processing the collection.
AbstractDocumentBuilder, line 285 (in search 4.1.1).  It then tries to
continue processing and then hits the same type of NPE that the
isExtraLazy check was doing.

I'm thinking the persistet.isExtraLazy line could be changed to include
a null check which would fall into the read() for this case and it would
work like it used to since I think you'd also hit this error if you
cleared a session and then tried to lazy load a collection.  I'll doctor
up a test in core and create a JIRA and all of that.

Shawn Clowater 

-----Original Message-----
From: hibernate-dev-bounces at lists.jboss.org
[mailto:hibernate-dev-bounces at lists.jboss.org] On Behalf Of Shawn
Clowater
Sent: Friday, September 14, 2012 10:23 AM
To: Sanne Grinovero
Cc: Hibernate
Subject: Re: [hibernate-dev] Regression after upgrade to 4.1.7.Final

I'm going to be rolling up to 4.1.7 sometime today since we just started
our new release and I'm trying to stay as current as I can.  I'll see
what I can see however, unless the search code is using nested sessions
then the patch wouldn't affect it.

Since you're already there, can you tell me if it's getting into the 

beforeTransactionCompletion( transaction ); call in the
transactionObserver from SessionImpl.  If not, then my patch comment is
just a red herring.

Shawn Clowater 

-----Original Message-----
From: Sanne Grinovero [mailto:sanne at hibernate.org] 
Sent: Friday, September 14, 2012 9:56 AM
To: Shawn Clowater
Cc: Steve Ebersole; Guillaume Smet; Hibernate
Subject: Re: [hibernate-dev] Regression after upgrade to 4.1.7.Final

I could reproduce the same problem by just running the Hibernate
Search testsuite, from master just changing the hibernate-core
dependency from 4.1.6 to 4.1.7

The failing test is
org.hibernate.search.test.embedded.EmbeddedTest

in the "hibernate-search-orm" Maven module, and reproduces exactly the
same stacktrace as reported by Guillaume.

In this case we are in a "before transaction completion" phase,
iterating over a previously non-initialized collection.

As you say, it's not finding the CollectionEntry; it would be very
nice if you could review your patch using this test to debug?

Sanne

On 10 September 2012 17:55, Shawn Clowater <shawn.clowater at entero.com>
wrote:
> Guillaume,
>
> Are you using any type of nested sessions in your code?  Steve applied
a
> patch that I created that now properly triggers the before transaction
> events in that case now (I had an issue with Search and it blowing an
> exception around sealing queues w/o it).  This may be why you're now
> seeing an issue in 4.1.7 and not 4.1.6.
>
> I'm not looking at the latest code but in 4.1.4 it looks like the
> getLoadedCollectionOwnerOrNull can blow a NPE if the CollectionEntry
> wasn't found but a test case is worth a thousand words.
>
> Shawn Clowater
>
> -----Original Message-----
> From: hibernate-dev-bounces at lists.jboss.org
> [mailto:hibernate-dev-bounces at lists.jboss.org] On Behalf Of Steve
> Ebersole
> Sent: Monday, September 10, 2012 8:53 AM
> To: Guillaume Smet
> Cc: Hibernate
> Subject: Re: [hibernate-dev] Regression after upgrade to 4.1.7.Final
>
> We are going to need a test (ideally without Search) reproducing this.
>
> On 09/10/2012 03:28 AM, Guillaume Smet wrote:
>> Hi,
>>
>> We just upgraded to Hibernate 4.1.7.Final (from 4.1.6.Final) and we
>> have a pretty bad regression with 4.1.7 and Hibernate Search 4.1.1.
>>
>> We have the following stack trace when saving an object:
>> Caused by: org.hibernate.HibernateException: Error while indexing in
>> Hibernate Search (before transaction completion)
>>       at
>
org.hibernate.search.backend.impl.EventSourceTransactionContext$Delegate
>
ToSynchronizationOnBeforeTx.doBeforeTransactionCompletion(EventSourceTra
> nsactionContext.java:186)
>>       at
>
org.hibernate.engine.spi.ActionQueue$BeforeTransactionCompletionProcessQ
> ueue.beforeTransactionCompletion(ActionQueue.java:662)
>>       at
>
org.hibernate.engine.spi.ActionQueue.beforeTransactionCompletion(ActionQ
> ueue.java:307)
>>       at
>
org.hibernate.internal.SessionImpl.beforeTransactionCompletion(SessionIm
> pl.java:607)
>>       at
>
org.hibernate.engine.transaction.internal.jdbc.JdbcTransaction.beforeTra
> nsactionCommit(JdbcTransaction.java:105)
>>       at
>
org.hibernate.engine.transaction.spi.AbstractTransactionImpl.commit(Abst
> ractTransactionImpl.java:175)
>>       at
> org.hibernate.ejb.TransactionImpl.commit(TransactionImpl.java:75)
>>       at
>
org.springframework.orm.jpa.JpaTransactionManager.doCommit(JpaTransactio
> nManager.java:512)
>>       ... 47 more
>> Caused by: java.lang.NullPointerException
>>       at
>
org.hibernate.engine.internal.StatefulPersistenceContext.getLoadedCollec
> tionOwnerOrNull(StatefulPersistenceContext.java:851)
>>       at
>
org.hibernate.event.spi.AbstractCollectionEvent.getLoadedOwnerOrNull(Abs
> tractCollectionEvent.java:75)
>>       at
>
org.hibernate.event.spi.InitializeCollectionEvent.<init>(InitializeColle
> ctionEvent.java:36)
>>       at
>
org.hibernate.internal.SessionImpl.initializeCollection(SessionImpl.java
> :1799)
>>       at
>
org.hibernate.collection.internal.AbstractPersistentCollection$4.doWork(
> AbstractPersistentCollection.java:524)
>>       at
>
org.hibernate.collection.internal.AbstractPersistentCollection.withTempo
> rarySessionIfNeeded(AbstractPersistentCollection.java:212)
>>       at
>
org.hibernate.collection.internal.AbstractPersistentCollection.initializ
> e(AbstractPersistentCollection.java:520)
>>       at
>
org.hibernate.collection.internal.AbstractPersistentCollection.read(Abst
> ractPersistentCollection.java:125)
>>       at
>
org.hibernate.collection.internal.PersistentBag.iterator(PersistentBag.j
> ava:266)
>>       at
>
org.hibernate.search.engine.spi.AbstractDocumentBuilder.appendContainedI
> nWorkForInstance(AbstractDocumentBuilder.java:296)
>>       at
>
org.hibernate.search.engine.impl.WorkPlan$PerEntityWork.processContained
> In(WorkPlan.java:525)
>>       at
>
org.hibernate.search.engine.impl.WorkPlan$PerClassWork.processContainedI
> nAndPrepareExecution(WorkPlan.java:297)
>>       at
>
org.hibernate.search.engine.impl.WorkPlan.processContainedInAndPrepareEx
> ecution(WorkPlan.java:144)
>>       at
>
org.hibernate.search.backend.impl.WorkQueue.prepareWorkPlan(WorkQueue.ja
> va:135)
>>       at
>
org.hibernate.search.backend.impl.BatchedQueueingProcessor.prepareWorks(
> BatchedQueueingProcessor.java:71)
>>       at
>
org.hibernate.search.backend.impl.PostTransactionWorkQueueSynchronizatio
> n.beforeCompletion(PostTransactionWorkQueueSynchronization.java:86)
>>       at
>
org.hibernate.search.backend.impl.EventSourceTransactionContext$Delegate
>
ToSynchronizationOnBeforeTx.doBeforeTransactionCompletion(EventSourceTra
> nsactionContext.java:183)
>>       ... 54 more
>>
>> I don't know if it's something obvious or if you need a self
contained
>> test case to reproduce it.
>>
>> I haven't created a JIRA because I don't know if it's an
>> incompatibility between Hibernate Search 4.1.1 and Hibernate 4.1.7 or
>> a real bug (either in HHH or HSEARCH).
>>
>> Feel free to ping me for more information or any further action.
>>
>> Thanks.
>>
>
>
> --
> steve at hibernate.org
> http://hibernate.org
>
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/hibernate-dev

_______________________________________________
hibernate-dev mailing list
hibernate-dev at lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hibernate-dev