On 05/05/2010 09:27 AM, Thomas Mortagne wrote: > On Wed, May 5, 2010 at 01:20, Sergiu Dumitriu<[email protected]> wrote: >> On 05/04/2010 02:57 PM, Ludovic Dubost wrote: >>> Le 04/05/10 10:40, Denis Gervalle a écrit : >>>>> - After a restart, I also got some "C" status, which are >>>>> not documented, >>>>> and I imagine, means conflict. But since this is just a >>>>> restart that cause >>>>> them, this not expected. Looking at the details, this >>>>> append on groups, >>>>> because of object GUID changes without any other changes, >>>>> and it may be >>>>> unrelated to your application in particular. >>>>> >>>>> Interesting. How come would GUID change in groups ? >>>>> >>>>> >>>>> I do not investigate, just noticed that after a tomcat restart, >>>>> one GUID has changed on all groups. >>>> Maybe the GUID is not set before. This is something to look at. >>>> >>>> >>>> No these were set and change. I have look closely to it, and it seems >>>> that all member object of each group has changed their GUID between my >>>> initial commit of those groups and the tomcat restart. >>> >>> Maybe the GUID specialists (Fabio, Sergiu) might have an idea ? >> >> Are only group objects affected? Or any other objects? >> >> Normally GUIDs are persisted for all non-custom-mapped objects. The only >> cases where GUIDs could change that I can think of are: >> >> - a deprecated hbm file which doesn't contain the mapping for guid >> - custom mapped objects which somehow don't include the mapping for guid >> - tags which sometimes don't really exist, although they mysteriously >> appear each time the document's objects are listed > > This should not happen anymore (the tags thing). >
Yes, but I doubt that anybody is using a trunk snapshot... -- Sergiu Dumitriu http://purl.org/net/sergiu/ _______________________________________________ devs mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/devs
