On Wed, Sep 06, 2000 at 07:32:22AM -0500, Brekke, Jeff wrote:
> > I'm +1 on #1. Do it right. Future changes to connection pooling
> would
> > require us to move to a service interface anyway. One could reverse
> #2 and
> > write a DBBroker implementation that uses the new service for
> comaptibility
> > ?
I'm +1 on moving to a service for both pool and map.
> > On an aside, I'm still not clear on why static data isn't destroyed
> within
> > the context on a restart/reload. I did verify that my jars are using
> the
> > context classloader and not the system classloader. I'm suprised more
> > haven't run into this problem before: eventually you will run out of
> > connections. Is there other static data that is not being destroyed
> on a
> > turbine restart? Are the services left hanging now also since we
> don't have
> > this destroy hook?
The thing is the VM doesn't restart, so nothing actually gets destroyed.
After a context reload, everything loaded by the old context classloader
is just left around for the garbage collector to pick up. But until it
does the objects are still there.
> >>> Second, regardless of the answer to the first question, shall I
> commit
> >>> the destroy() chaining (Turbine servlet's destroy() method iterates
> > its
> >>> Services, calling destroy on each)?
> >
> >> +1 from me, but you need two more.
> >
> > +1
+1 - sounds sensible
--
Sean Legassick
[EMAIL PROTECTED]
homo sum: humani nihil a me alienum puto
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Search: <http://www.mail-archive.com/turbine%40list.working-dogs.com/>
Problems?: [EMAIL PROTECTED]
