hi cédric On 11/10/05, Cédric Damioli <[EMAIL PROTECTED]> wrote: > Hi all, > > I'm playing with multi-threaded stuff (SVN trunk from less than 10 > minutes). > > Let suppose that two different threads acquire a Session, then get the > same Node X and then add a new child to this Node, and then call the > save() method. > > First save() call will success, and second one will lead to : > javax.jcr.InvalidItemStateException: > b13b8065-5115-4bd3-87cf-5544879209e8: the item cannot be saved because > it has been modified externally. > The given UUID being the actual UUID of the Node X. > I assume this behaviour is ok, isn't it ?
correct. > > But, second step, just before calling the method save(), I call > session.refresh(true), and what I'm getting is not what I was expecting : > firstly, the InvalidItemStateException is here again > and secondly, after debugging, it appears that the refresh(true) doesn't > behave as expected : after the second call to refresh(true), the second > Sessions does not know (according to itemExists()) the first child > created by the first Session. > > My code is equivalent to : > > Node X = session.getNode(path); > X.addNode(whateverUniqueName); > session.refresh(true); > session.save(); > > > Do I misunderstood the refresh() method ? Or is there a bug here ? the current behaviour of refresh=true is according to the spec. the spec does not mandate to merge changes made by different sessions on the *same* item, in your case parent node. cheers stefan > > Regards, > > -- > Cédric Damioli > ANYWARE TECHNOLOGIES > Tel : +33 (0)5 61 00 52 90 > Fax : +33 (0)5 61 00 51 46 > http://www.anyware-tech.com > >
