Ya, me and Ralf talked a bit about it last night and have a quick test case to raise the issue fairly easily. Will create a new bug this morning and post the info.
It comes down to when we dicover that the object we are loading is actually an extended object and when we let the rest of the world (the transaction) know about it. -Nick On 8/11/05, Werner Guttmann <[EMAIL PROTECTED]> wrote: > Nick, > > actualy, can you please raise a new bug with the Castor Jira instance ? > > Werner > > -----Ursprüngliche Nachricht----- > Von: Nick Stuart [mailto:[EMAIL PROTECTED] > Gesendet: Mi 10.08.2005 19:46 > An: dev@castor.codehaus.org > Cc: > Betreff: Re: [castor-dev] NPE with extended classes. > > > > Ok, got it to fail using a very simply test case. Attached the patch > to the current test case as well as the extra object I created. If you > run it you should get the following exception: > org.exolab.castor.jdo.PersistenceException: Requested to load/fetch an > object of type jdo.test1158.ExtendedObject, where persistent storage > returned an object of type class jdo.test1158.BaseObject > > The problem is that it sees the parent of DependentObj as being loaded > because BaseObject is loaded. And when it goes to try and load the > ExtendedObject it just gets all confused. Not sure what the design for > this is, but it looks like you are loading BaseObject seperatly first, > then, if you find an extended object if you find a class that extends > BaseObject with the same ID and replacing it? Not sure one this one. > > -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] -------------------------------------------------
