Ok, good to hear, and thats exactly what I was doing (call getDatabase
many times).  Also, for a quick patch so that  I could continue on
working on my end was to simply pass "local" to 
getTransactionManagerFactory()

It works for me, but I know it's not 'correct' or that safe.

-Nick


On 8/9/05, Ralf Joachim <[EMAIL PROTECTED]> wrote:
> Hi Nick,
> 
> when working at CASTOR-1158 I recognized myself that we introduced a
> problem with transactions at commit of CASTOR-1173. The problem apears
> if you call getDatabase() multiple times on the same JDOManager instance.
> 
> The patch of CASTOR-1158 includes a quick fix that solves the problem
> for this test. When commiting CASTOR-1158 I didn't commit this quick fix
> as I thought it may introduce others and I just started to work at
> CASTOR-1179 which is a bigger refactoring of the transaction handling.
> I'm about half the way gone to resolve CASTOR-1179 but have no results
> to publish yet.
> 
> Ralf
> 
> 
> Nick Stuart schrieb:
> > Hey guys, just trying to check out the latest CVS build and can't get
> > it to fire up. It looks like the TransactionManager name is getting
> > lost somewhere when I try to load a database. Here's the stack trace
> > from the castor side of things.
> >
> > java.lang.NullPointerException
> >       at java.util.Hashtable.get(Hashtable.java:336)
> >       at 
> > org.exolab.castor.jdo.transactionmanager.TransactionManagerFactoryRegistry.getTransactionManagerFactory(TransactionManagerFactoryRegistry.java:109)
> >       at org.exolab.castor.jdo.JDOManager.getDatabase(JDOManager.java:803)
> >
> > Also note, with the stable release of 0.9.7 this exact mapping worked
> > fine. I dont know whats missing as it seems to be kind of hard for the
> > local var 'transactionMode' to not get set and NOT throw an exception,
> > but its happening somewhere.
> >
> > -Nick
> >
> > -------------------------------------------------
> > If you wish to unsubscribe from this list, please
> > send an empty message to the following address:
> >
> > [EMAIL PROTECTED]
> > -------------------------------------------------
> 
> -------------------------------------------------
> If you wish to unsubscribe from this list, please
> send an empty message to the following address:
> 
> [EMAIL PROTECTED]
> -------------------------------------------------
> 
>

-------------------------------------------------
If you wish to unsubscribe from this list, please
send an empty message to the following address:

[EMAIL PROTECTED]
-------------------------------------------------

Reply via email to