I was checking bugzilla before reporting these bugs. Like other opensource
projects, it is difficult to evaluate the cutting edge of the release process in
any meaningful way. It's a moving target. I'm just a little surprised that such
obvious problems made it through to the "2.0 Final" release. Hopefully, when you
fixed these bugs, the causes were added to a test harness.
I will look at 2.1. Thanks for the heads up.
jim
----- Original Message -----
From: "Tom Cook" <[EMAIL PROTECTED]>
To: "JBoss-User" <[EMAIL PROTECTED]>
Sent: Thursday, March 08, 2001 5:44 PM
Subject: Re: [jBoss-User] Bug: Synchronization error using commit option A
> You are reporting lots of bugs against 2.0 final - the code base has moved
> considerably since then. You really need to get a more recent version.
>
> Tom
>
> On Thu, 8 Mar 2001, James Cook wrote:
>
> > In our ejbLoad we are all accustomed to calling:
> >
> > PK key = (PK)this.entityContext.getPrimaryKey();
> >
> > We do this because the bean's attributes are not yet set and the container
will
> > use the entitycontext to pass the pk.
> >
> > In ejbRemove(), we do not have to resort to this step because the bean's
> > attributes will be set with a call to ejbLoad() in the case of commit
options B
> > and C.
> >
> > But how do we execute ejbRemove() when using commit option A? When using
commit
> > option A, the bean is *not* synced with the db because ejbLoad() does not
fire.
> > The container however must ensure that you get the appropriate bean instance
> > that matches the PK you wish. So if I do:
> >
> > MyRemote remote = getMyHome().findByPrimaryKey(pk);
> > remote.remove();
> >
> > the FBPK method should return a remote reference to the bean in the ready
pool
> > that matches the pk value I pass in, if it exists already. If it does,
ejbLoad()
> > is not called. However, when I invoke remote.remove(), the pk attribute of
the
> > bean is null!? This appears to be a bug.
> >
> > jim
> >
> >
> >
> >
> >
> > --
> > --------------------------------------------------------------
> > To subscribe: [EMAIL PROTECTED]
> > To unsubscribe: [EMAIL PROTECTED]
> >
> >
>
> --
> "If you mess with something for long enough it will break." - Schmidt
>
>
>
> --
> --------------------------------------------------------------
> To subscribe: [EMAIL PROTECTED]
> To unsubscribe: [EMAIL PROTECTED]
>
>
--
--------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
