Hi Gul,

On Nov 22, 2007, at 10:18 AM, Gul Onural wrote:


Sorry, I should have said GregorianCalendar.

What we do in our product is that we retrieve the objects from the
database through JPA, and construct another set of objects(business
objects) from these JPA objects. So, our JPA objects are not directly
passed to the business logic layer. At the time the business objects are used the JPA objects are already detached and context closed. Therefore when we want to do anything with the Calendar that we retrieved through
JPA, the openjpa gives us an exception saying context closed.

This is probably a bug. Once the Calendar instance is detached, it should behave normally, just like a new instance of GregorianCalendar (it should implement all of the behavior of GregorianCalendar). If you have a small reproducible test case, please file a JIRA.

May be
better thing to do was to clone the Calendar JPA returns and pass this
clone as our business object to the other layers of the product.

I'd hate to see you go through all this trouble if we can make the OpenJPA Calendar do what you need it to do.

Craig

Gul


-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent: Thursday, November 22, 2007 1:08 PM
To: [email protected]
Subject: Re: Calendar problem in openjpa

Hi Gul,

The abstract class Calendar doesn't have a concrete add method either.
It's the GregorianCalendar that implements the abstract method.

Are you saying you want a GregorianCalendar as the field type of your
persistent Calendar? Have you tried declaring that as the field type?

And why don't you want the openjpa implementation of Calendar?

Regards,

Craig

On Nov 22, 2007, at 9:44 AM, Gul Onural wrote:


We discovered a strange behavior about the usage of Calendar and I was

wondering if this was expected :

There is a Calendar field in one of our JPA objects. If I try to use
this field through its getter method, the instance I got back is the
open jpa implementation of the Calendar interface, not Sun
implementation (i.e. not java.util.Calendar). We discovered this when
we try to call "add" on this implementation which openjpa
implementation doesn't support. We obviously do not want t use openjpa

implementation of the Calendar in our code.

I was wondering if it is the right behavior ? Shouldn't openjpa give
java.util.Calendar back when it unserializes the Calendar field ?

Gul


Craig Russell
Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
408 276-5638 mailto:[EMAIL PROTECTED] P.S. A good JDO? O, Gasp!


Craig Russell
Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
408 276-5638 mailto:[EMAIL PROTECTED]
P.S. A good JDO? O, Gasp!

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to