Actually I am doing the clear, once I got rid of it the exception disappeared. SO IMHO no bug
Phill -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Craig Russell Sent: April 30, 2007 5:08 PM To: open-jpa-dev@incubator.apache.org Subject: Re: How to debug object retrieveal in JPA? IMHO, this is a bug in the generated code for pcNewInstance. I don't understand why the pcNewInstance should ever call pcClearFields. It just got finished creating a new instance, and it knows that all fields have been set to their Java default values. [That's why we require the user to specify a no-args constructor, so the implementation doesn't have to initialize anything.] I don't think that calling pcClearFields is warranted during this life cycle phase. Craig Caused by: java.lang.NullPointerException at java.util.ArrayList.addAll(ArrayList.java:472) at ca.BidSpec.emall.user.Person.pcsetPhoneNumbers(Person.java:727) at ca.BidSpec.emall.user.Person.pcClearFields(Person.java) at ca.BidSpec.emall.user.Person.pcNewInstance(Person.java) at org.apache.openjpa.enhance.PCRegistry.newInstance(PCRegistry.java:121) I had read this and assumed that it was an OpenJPA-generated method, but if so, there shouldn't be a line number. Did you reverse-compile your classes at some point and change the synthetic interception methods? That is the original method with a new name. As the documentation states, the getters and setters used in property access may appear with a "pc" prefix in stack traces under OpenJPA.
