On Tue, Apr 28, 2009 at 2:10 AM, Yoav Landman <[email protected]> wrote: > Well, I found a good discussion about the underlying logic in > https://issues.apache.org/jira/browse/JCR-2011 :) > However, in my case, none of the tx changed the mixins that are applied to > the node in question. So, does that mean that any concurrent changes to a > node that has mixins on it are deemed as conflicting? Isn't that a bit too > harsh or am I missing something?
Just to be sure: What do you mean by concurrent change? In different transactions or by concurrent access to the same session? Otherwise I'd think a content-based diff sounds better than to check the references. But I have not much knowledge about this code; it might be that there should always be only one instance of the NameSet object for a given name, so that a reference check would work. This could indicate a different problem. Anyone? Regards, Alex -- Alexander Klimetschek [email protected]
