On Thu, Sep 12, 2002 at 11:18:48AM +0200, Max Rydahl Andersen wrote: > > Hi, > > On Thu, Sep 12, 2002 at 09:11:09AM +0200, Max Rydahl Andersen wrote: > > > This is how I at some point solved the problem in another persistence > layer. > > > This works "great" as long as one remember to cleanup the cache after > the > > > transaction is completed. > > > If this is not done the cache can contain "leftovers" from the previous > > > thread that might > > > not be commited (or does OJB handles this ?) > > > > Right but I think this happens with the OJB single-cache strategy, too. > > If i change an Object held in the cache and don't do a store (and it is > > not stored via auto-update) it is left changed but not committed in the > > cache. > > I really, really, really wonder how this can be sound in a production > environment ?! > Why have this not failed constantly ? Is it just not happening in the > realworld ?
I was talking about using PersistenceBroker API only. Sorry I forgot to mention. I think here it is simply up to the programmer to track which objects he modifies/stores/removes from cache. If you use the ODMG layer, it does the work of committing all changed objects (since by write-locking an Object it is registered for later update to the db) Jens -- Jens Krämer [EMAIL PROTECTED] -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
