Timur Mehrvarz wrote: > > On 21.05.2007, at 21:26, Bob Scheifler wrote: > >>> Done. Now, when I experience a notification being dropped, a >>> NoSuchObjectException is indeed being logged: >> >> So first, I had things bass ackwards, you really wanted to turn >> on client remote call logging in outrigger, not server remote call >> logging. And the logging change previously discussed should not >> have altered seeing the logging record you provided, as it's using >> an outrigger logger, not an RMI one. Are you sure something else >> hasn't changed? >> >>> Obviously, my JavaSpace client is becoming "permanently unavailable >>> (NoSuchObjectException)" and "further events will not be delivered to >>> it." Is anyone able to speculate, how my humble client may cause this >>> behavior? >> >> If your client isn't holding a strong reference to the listener impl, >> it can get GC'd. >> >> But this exception doesn't correlate with your description of >> occasionally dropped events. Or is your client re-registering >> for events, and you didn't explain that before? Or are you >> perhaps seeing a new different problem than before? > > I'm afraid, it's still the same problem. The polling is good to have, as > a backup. But right now, my app depends on it. It wouldn't work without > it. I need to stop the NoSuchObjectExceptions. I do re-register for > every event. I am holding strong references to the listeners. As far as > I can see, their references do not get GC'd. What else could trigger > Outrigger to throw NoSuchObjectExceptions on occasion? >
Can you explain more about what you mean by re-register for each event? Can you maybe post the related code for that. > I'd like to look at the code handling this. Do you have a pointer? > Not sure I understand what you're asking for here - are you talking about your code or the Outrigger code? > Timur > > Hope that helps, Dan.
