Jon Stevens wrote:
>
> on 9/1/2000 12:12 PM, "Paul O'Leary" <[EMAIL PROTECTED]> wrote:
>
> > I don't have time to run this down today but this behavior is most likely
> > caused by DBBroker's singleton implementation.
> >
> > DBBroker.getInstance caches a global instance of DBBroker in a class static
> > variable. Since I assume that DBBroker is loaded as a system class, unless
> > the VM instance is cycled you'll always be looking at the same copy of the
> > class info and, therefore, the same static variable.
> >
> > If this is a serious problem, one solution may be to have an explicit
> > tear-down of the connection pool when Turbine shuts down. Probably want to
> > do this anyway because, as I noted in a previous mail, the pool currently
> > shuts itself down in DBBroker's finalize() method which is a bad idea.
> >
> > I'll have some more time soon to look at this in more detail...
> > cheers,
> > PaulO.
>
> That is a very good point. My advice is to ALWAYS load Turbine in the
> context classloader, not the system class laoder.
I agree, but worry about people not knowing enough to differentiate
between the two. Jon, if you think that implementing the shutdown would
be straightforward, I would like to go ahead with that rather than see
this come up repeatedly. ;)
Dan
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Search: <http://www.mail-archive.com/turbine%40list.working-dogs.com/>
Problems?: [EMAIL PROTECTED]