On Thu, 3 Oct 2002 [EMAIL PROTECTED] wrote:
>
>
> ----- Original Message -----
> From: Eivind Waaler <[EMAIL PROTECTED]>
> Date: Thursday, October 3, 2002 9:52 am
> Subject: [Middlegen-user] RE: Double PK field trouble with EJB CMP2.0
> composite PK in CMR/CMP clash
>
> > Allright, I've deployed the airline sample on Weblogic 7.0, SQL Server
> > 8. Seems like we actually get the same problem there. I added the
> > following code to the InsertReservationAction:
> >
> > PersonLocal personLocal = ...
> > FlightLocal flightLocal = ...
> >
> > ReservationLocal reservation = ReservationUtil.getLocalHome().create(
> > reservationForm.getRegistrationUtc(),
> > reservationForm.getComment(),
> > personLocal,
> > flightLocal
> > );
> >
> > Testing of this on Weblogic ends in the following:
> >
> > <03.okt.02 10:29:17 CEST> <Error> <HTTP> <101046>
> > <[ServletContext(id=3401671,name=airline-web.war,context-
> > pathActionServlet: Create error
> > javax.ejb.CreateException: In EJB Reservation, primary key field
> > 'flightIdFk' was not set during ejbCreate. All primary key fields
> > must be
> > initialized during ejbCreate.
> > at
> > airline.ejb.ReservationCMP_1qdy24__WebLogic_CMP_RDBMS.pkCheck
> (ReservationCMP_1qdy24__WebLogic_CMP_RDBMS.java:2207)
> > at
> > airline.ejb.ReservationCMP_1qdy24__WebLogic_CMP_RDBMS.ejbCreate
> (ReservationCMP_1qdy24__WebLogic_CMP_RDBMS.java:2259)
> > at java.lang.reflect.Method.invoke(Native Method)
> > at weblogic.ejb20.manager.DBManager.create(DBManager.java:737)
> > at
> > weblogic.ejb20.manager.DBManager.localCreate(DBManager.java:716)
> > at
> > weblogic.ejb20.internal.EntityEJBLocalHome.create
> (EntityEJBLocalHome.java:182)
> > at
> > airline.ejb.ReservationCMP_1qdy24_LocalHomeImpl.create
> (ReservationCMP_1qdy24_LocalHomeImpl.java:97)
> > ...
> >
> > I'm sure we had this bean working som months ago on the cvs
> > version of
> > middlegen at the time. That was however WLS6.1 I used back then.
> >
> > Is this a Middlegen problem, xdoclet problem or weblogic 7.0
> > problem?? Anyone got it working? I just can't see what's wrong
> > with this
> > bean...
> >
>
>
> I don't know what the EJB spec says about this. Anyone? It's definitely
> not an XDoclet problem. It's a WLS problem, but maybe Middlegen can
> work around it?
>
> What if the ejbCreate methods that take CMR/CMP fields as this example
> does the following:
>
> ejbCreate(..., PersonLocal person, ...) {
> setPersonIdFk( person.getPersonId() );
> }
>
> -in addition to or in stead of calling the CMR setter in ejbPostCreate
> ()? Eivind, can you try that?
Yes this works. The CMR setters in ejbPostCreate() must be removed,
otherwise you get an error since these are read only when part of the
primary key.
I submitted a bug report as I'm not entirely sure how to fix this the best
way. Also don't know if this works with JBoss..
.eivind
>
> I don't know if JBoss or other containers would like this, so maybe
> we're facing app-server-specific java code here. We should try to avoid
> that, but if it's the only solution, let's go for it.
>
> /Aslak
>
> > .eivind
> >
> > snippysnippy
> > >
> > > So it's *exactly* like the airline sample, then. Did you look
> > at that?
> >
> > Yes, it probably is very much like the airline example. I've
> > discussedthis
> > with Eivind today, and we figured out it would probably be a
> > good idea
> > to
> > try to deploy the airline app to weblogic, and see if the problem
> > reincarnates there. Placed on personal TODO list, though.
> >
> > >
> > >> Result is that entity bean for B gets FK/PK fields acting both
> > as CMR
> > and CMP fields. Rules are approx. as follows: 1) a CMR field is
> > snip snip snip>
> >
> > > Create A and C first and then create B ;-)
> > > -or ask on http://newsgroups.bea.com/
> >
> > This is actually what is done, and it has been verified that A
> > and C both
> > are assigned valid id's from autogenerator, so the problem is
> > that the
> > ids
> > (PK's) are never transferred to the PK fields of the PK class of
> > bean C,
> > probably due to the CMR/CMP confusion.
> >
> > However, I got around the problem by giving the B table its own
> > autonumber
> > PK, though conceptually not neccessary. As a consequence, my ORM
> > model is
> > somewhat messed up. I can live with that for now, though.
> >
> >
> >
> > Regards,
> >
> >
> > Kristoffer Moe
> > Senior Consultant
> > Elan It ReSource
> >
> >
> >
> >
> > -------------------------------------------------------
> > This sf.net email is sponsored by:ThinkGeek
> > Welcome to geek heaven.
> > http://thinkgeek.com/sf
> > _______________________________________________
> > middlegen-user mailing list
> > [EMAIL PROTECTED]
> > https://lists.sourceforge.net/lists/listinfo/middlegen-user
> >
>
>
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
middlegen-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/middlegen-user
