Mailing-List: contact user-help@geronimo.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@geronimo.apache.org
Received-SPF: pass (herse.apache.org: local policy)
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Subject: AW: AW: Problems using openJPA
Date: Wed, 18 Jul 2007 10:01:49 +0200
Message-ID: <C40B585EBE11B54CBF47550F01D5AC8901DB8AD6@exchange.dbh.de>
In-Reply-To: <EBE8FB8C-0375-479D-8482-06B81EB3CD32@yahoo.com>
Thread-Topic: AW: Problems using openJPA
Thread-Index: AcfIub+SAvQboJkgTf2OtRF3lsQyZAAUoj0g
From: "Ueberbach, Michael" <michael.ueberbach@dbh.de>
To: <user@geronimo.apache.org>

Hello David,

To your last question: The SUN spec says=20

"In addition to returning and setting the persistent state of the =
instance, the property accessor methods
may contain other business logic as well, for example, to perform =
validation. The persistence provider
runtime executes this logic when property-based access is used."

openJPA follows this in offering the two possibilities of accessing. The =
openJPA manual says

"JPA offers two modes of persistent state access: field access, and =
property access.
To use property access, set your entity element's access attribute to =
PROPERTY, or place your metadata and mapping annotations on the getter =
method.
Each class must use either field access or property access for all =
state; you cannot use both access types within the same class."

The whole problem is that the described way to choose the property =
access by placing the annotations at the getter methods does not show =
any effects.
More worse openJPA throws an org.apache.openjpa.util.UserException (as =
described) and does not choose any of the two alternative ways.

This exceptions sounds as if the annotations set at the getter methods =
were not reckognized when parsing the entity classes. The question for =
me is:
Have I to do something more (or have I chosen the wrong annotations or =
what ever) to get this code working?

The test with jboss was done to see whether there also an error would be =
visible, but it was not. I agree that this gives no final security about =
the compliance of jboss. Indeed my aim is to get the thing working on =
geronimo/openJPA!

Thanks for any help
Michael =20


-----Urspr=FCngliche Nachricht-----
Von: David Jencks [mailto:david_jencks@yahoo.com]=20
Gesendet: Dienstag, 17. Juli 2007 23:30
An: user@geronimo.apache.org
Betreff: Re: AW: Problems using openJPA


On Jul 17, 2007, at 9:25 AM, Ueberbach, Michael wrote:

> Hi Jacek,
>
> I had to make a little break, now I'm trying to get back to the last=20
> situation.
>
> Following your last advices:
>
> reducing the persistance.xml to the neccessary elements works for=20
> deploying, but does not help in fetching data. In fact the same=20
> exception (...Could not locate metadata for the class ...) is thrown.
>
> I checked the code again:
> - no field is annotated
> - every getter is annotated (at least with @Basic)
> - every other method named with get... Is annotated @Transient
>
> Therefore the access type has to be clear (=3D PROPERTY). =20
> Nevertheless this is not recognized.
>
> So I've tried two things:
>
> First:
> I put a xml mapping  file (orm.xml) inside the deployment jar, where I =

> explicitly set the access type to PROPERTY for all classes.
> The deployment works, but the warning
>
> "<1.0.0-SNAPSHOT-SNAPSHOT nonfatal user error>
> org.apache.openjpa.util.UserException: Type=20
> "de.lama.sample3.stammdaten.ejb.PersonBean" attempts to use both field =

> and property access. Only one access method is permitted."
>
> remains and the method call still doesn't work.
>
> Second:
> I deployed the whole stuff unchanged (except the deployment plan) to a =

> jboss server(4.04) to compare the results. Here everything works as=20
> expected.
>
> As result it seems to me that if there isn't an error inside the=20
> deployed code that jboss does not find, there must be something wrong=20
> with parsing the classes annotations in geronimo resp. openjpa.

I haven't looked into your problem in detail, but I don't think you can =
conclude from hibernate's ability to deploy an app X that X is entirely =
spec compliant and can be deployed on all other compliant jpa =
implementations.

Out of curiousity why are you using property access?  I have yet to =
understand why anyone would do this given a choice.... hoping for =
enlightenment.

thanks
david jencks

>
>
> regards
> Michael
>
>
> -----Urspr=FCngliche Nachricht-----
> Von: Ueberbach, Michael [mailto:michael.ueberbach@dbh.de]
> Gesendet: Dienstag, 10. Juli 2007 18:06
> An: user@geronimo.apache.org; jacek@laskowski.net.pl
> Betreff: AW: Problems using openJPA
>
> Hi Jacek,
>
> Thanks for your hints. To the first point: I've already seen this in=20
> the spec. I'm quite shure there is no access via the instance=20
> variables (but I will prove it again). I will also try to eliminate=20
> the entries in persistance.xml that are not neccessary. I'll tell the=20
> results as soon as possible.
>
> Michael
>
> -----Urspr=FCngliche Nachricht-----
> Von: eljotpl@gmail.com [mailto:eljotpl@gmail.com] Im Auftrag von Jacek =

> Laskowski
> Gesendet: Dienstag, 10. Juli 2007 11:19
> An: user@geronimo.apache.org
> Betreff: Re: Problems using openJPA
>
> On 7/10/07, Ueberbach, Michael <michael.ueberbach@dbh.de> wrote:
>> Hello Jacek,
>>
>> meanwhile I'm able to send the jar I deployed to geronimo. I attach=20
>> it to this mail, sources should be included. Also the persistence.xml =

>> that is used.
>> As you can see I put @Basic annotations to all property getters=20
>> explicitly (although I think this should be default) to make shure=20
>> that access should be done by property (and not by field).
>> Nevertheless this seems not be recognized by openJPA. It would be=20
>> great if you have any idea why and what should be done.
>
> Hi Michael,
>
> I've just started looking into it, but thought I'd tell you that when=20
> you use property-based access (via getters) you should not access the=20
> entity state via instance fields.
>
> You don't need to specify vendor element in the persistence.xml file.
> Geronimo defaults to OpenJPA. According to the spec (6.2.1.4 provider
> p.135):
>
> The provider element must be specified if the application is dependent =

> upon a particular persistence provider being used.
>
> You don't need to specify class elements in Java EE env (as opposed to =

> Java SE env). The container looks for other files unless exclude-=20
> unlisted-classes element is specified (which is not the case in your=20
> example).
>
> See if that helps out. Going to spot other (mis)uses of the spec that=20
> might cause the trouble.
>
> Jacek
>
> --
> Jacek Laskowski
> http://www.JacekLaskowski.pl