Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [eclipselink-dev] bug 336731 <shared-cache-mode>ENABLE_SELECTIVE</shared-cache-mode>

Hi Gordon,


I have now proceeded as follows:


-          I have added an isolated test case to reproduce this issue:


eclipselink.jpa.wdf.test … org.eclipse.persistence.testing.tests.wdf.jpa1.inheritance.ChangeSubclass/test


-          In other tests, I have changed the behavior so that subclasses are not changed any longer.


Checked in at #9541.





From: Gordon Yorke [mailto:gordon.yorke@xxxxxxxxxx]
Sent: Montag, 30. Mai 2011 19:41
To: Goerler, Adrian
Cc: Dev mailing list for Eclipse Persistence Services
Subject: Re: bug 336731 <shared-cache-mode>ENABLE_SELECTIVE</shared-cache-mode>


I have updated the bug.

Goerler, Adrian wrote:

Hi Gordon,


I am currently cleaning up the bugs assigned to be. This one is still open:


Did you have a chance to look into this?


Best regards,





Adrian Görler

Pflichtangaben/Mandatory Disclosure Statements:


From: eclipselink-dev-bounces@xxxxxxxxxxx [mailto:eclipselink-dev-bounces@xxxxxxxxxxx] On Behalf Of Goerler, Adrian
Sent: Mittwoch, 9. Februar 2011 18:13
To: Gordon Yorke
Cc: Dev mailing list for Eclipse Persistence Services
Subject: [eclipselink-dev] bug 336731 <shared-cache-mode>ENABLE_SELECTIVE</shared-cache-mode>


Hi Gordon, other,



I have started an investigation to get the WDF test running with <shared-cache-mode>ENABLE_SELECTIVE</shared-cache-mode>.


I will track progress here:



The bug 336731 proposes some minor fixes and adjustments to the tests but also revealed an interesting observation:


In an entity hierarchy, some subclass with a particular id had been cached and is removed and evicted from the cache later. It should be completely gone now. However, it seems that upon eviction, the EclipseLink cache does not truly remove the subclass but just invalidates it and keeps a zombie object with the particular subclass. If I now create a different subclass with the particular id, it cannot enter the cache as the cache expects it to have the attributes of the former subclass.


I don’t think this is an issue in real life as objects don’t just change their subclass. But that the subclass information survives the removal of an object is kind of strange.


In 336731, I am proposing a change that adjust the tests so that this issue does not occur. But if you think it is worth investigating I can keep the tests as is to have a repro.


What do you think?




Adrian Görler

Pflichtangaben/Mandatory Disclosure Statements:




Back to the top