Re: AW: Bug fix

[Marcel Ruff <ruff at swand.lake.de>]

Konrad Krafft wrote:
> 
> > -----Ursprüngliche Nachricht-----
> > Von: ruff at swand.lake.de [mailto:ruff at swand.lake.de]
> > Gesendet: Mittwoch, 23. Februar 2000 13:02
> > An: Konrad Krafft
> > Betreff: Re: Bug fix
> >
> >
> > Konrad Krafft wrote:
> > >
> > > Hello,
> > >
> > > I fixed a problem with non existing keys in the xmlBlaster.
> > > Description:
> > > If you call a 'get' on an non existing key in the xmlBlaster, it throws
> > > a null pointer exception. This causes problems in my programs
> > > because my client only gets a CORBA NO_IMPLMENT Exception.
> > >
> > > solution:
> > > I fix this problem by throwing an XmlBlasterException.
> > > There might be better ways but this was the fastet solution.
> >
> > Hmm, thanks for finding/locating/fixing the bug.
> > But the philosophy in xmlBlaster is another:
> >
> > If you subscribe to a yet unknown message with an exact key-oid,
> > this subscription is accepted. If later a data source publishes
> > this message, you will get this updated.
> Oh yes, I remember this philosophie ;-)
> But I didn't use a subscription, I just want to get a special entry
> in the xmlBlaster-DOM.
> The difference is, a subscription generates no MessageUnit, a 'get' does.

Sorry, you are right, you are using a synchronous get().
Could you please commit your fix into cvs?

I extended the testuite to catch this case as well.


Marcel
> 
> Konrad
> >
> > So it is not important, *when* you subscribe, before or after
> > a data source published a message.
> >
> > This decoupling is very important in MOM.
> >
> > Marcel
> >
> >
> > >
> > > Konrad
> > >
> > >
> >
> > --
> > Marcel Ruff
> > ruff at swand.lake.de
> > http://www.lake.de/home/lake/swand/
> > http://www.xmlBlaster.org
> >

-- 
Marcel Ruff
ruff at swand.lake.de
http://www.lake.de/home/lake/swand/
http://www.xmlBlaster.org