On Fri, Feb 27, 2009 at 10:49 PM, Alex Shneyderman <[email protected]> wrote:
> I was waiting for the 0.6 hoping this will be fixed but not much > changed and there is still a bug in > > org.qi4j.index.rdf.assembly.RdfNativeSesameStoreAssembler > > it needs EntityTypeSerializer in addition to EntityStateSerializer (at > least in my case there were complaints by qi) Something fell thru the cracks obviously... > Should I just abandon my efforts of trying to make it work? How do you > guys persist? There are no tests doing native configuration. Well, the reason we 'persist' is that we consider "Assemblers" to be convenience, and there will probably never be assemblers for every case, so I think that many people (myself included) in the future will not rely heavily on provided assemblers, but often use either a basic one or even totally there own assembly information. Potentially, there might be generic assemblers that obtain the assembly information from other sources than programmatically, and so forth. > I should be able to > bring down jvm start a new one and be able to find my old friends. > WDYGT? Sure, the Native store Sesame indexer is what you currently need. I anticipate a lot of work in the indexing area before the 1.0 release, and it just haven't had enough priority so far. > Note, also that the above will work and the data is still there on the > consequtive runs but I have to update at least one entity to > recalculate in-memory index. So, maybe I can nudge the indexer to > re-index somehow? Any code sample would be appreciated. Re-indexing is one of the things that MUST be fixed before a 1.0 release. For instance, JVM crashes between the "store" and "index" would create an invalid index, and such scenario must be recoverable. IIRC, that is why the EntityStore has contract to return an Iterator for all Entities. > If there is an alternative way to search for entities, I would be > happy to hear about them. Help us implement Indexing!! :-) You have a starting point (existing code), we will assist to answer any question, even by IM if that helps, and you have commit rights... From our perspective a good alternative ;-) > Alternatively maybe someone could take a look at what I got and see if > there is something really obvious that's incorrect (I could zip my > sample up and send it to anyone who might be interested). Sure, pass it my way, and I will try to find time to take a look during the weekend. Cheers Niclas -- http://www.qi4j.org - New Energy for Java _______________________________________________ qi4j-dev mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/qi4j-dev
