Akil wrote: > I am a newbie at both WebDAV and Slide, though I have read most of the > documentation as well as much of the API JavaDocs for Slide as well as some > of the specification documents at webdav.org. I also looked through the > archives of the user and dev lists. That should give you an idea of my level > of knowledge. > > I started a project a couple of weeks back to facilitate the use of the > SourceWare Server as the content indexer with Slide in order to be able to > provide full text search. Eventually I plan to implement a SourceWare Store. > (SourceWare Server is a database with an integrated full text search engine > supporting Batch and Incremental Indexing, Flexible Semantics, Field > searches, ... with a Java based CORBA API i.e. no SQL.) > I will be setting up a web page for this project in a week or so. > > In the short term, I am trying to just do the content indexer piece in > SourceWare. The Search piece will be Tomcat based, access the indexing > SourceWare server directly and act as a client of Slide when retrieving > documents. > > There are a number of ways to implement a SourceWare based content indexer > for Slide and the most appropriate way would be to > implement the IndexStore interface and associated mechanisms. > > Due to the urgency of the situation, I would like to take a short-cut for > the first pass and implement it by adding a ContentInterceptor that has a > poststore and a postremoval method. I have a few questions regarding this > approach. > > 1. I saw some information regarding the addition of the > post-removal method to the ContentInterceptor class (as well as associated > mechanisms) in the dev list archive. How come the code is not in > last-night's build? Is it going through a code review process? Or have I > missed something?
I've sent a patch to add a postRemoveContent() method to the ContentInterceptors but this one has to be tested by commiters before to be included into cvs tree (it modifies some critical classes like StoreImpl). > 2. Has the issue regarding access to the Namespace > Configuration in the ContentInterceptor class been resolved? In the patch too, but I think it could be done smarter than it is. > 3. Although, this is not required in the short term, > eventually, the Search Engine would also need the Access Control List > information per node to be able to tune the search for a particular user. > Post search verification is not feasible when the visible portion of the > namespace is sparse for a given user. > > For this purpose, I would need to get the list of > permissions that are explicitly set for a new node. Moreover, I would need > to intercept any changes in the permissions for an existing node. Any ideas > on this? The way I understood it, the content interceptor is only invoked > when the node being added/retrieved/removed has non-zero length content. How > do I intercept security permission creations/modifications/deletions. To retrieve/set permissions on a given node, you can use the security helper, but I don't know how you could simply intercept permissions changes (patching SecurityImpl ?). > 4. Any major flaw in my approach to either the short term / long > term projects? Personnaly, I think that indexing is a heavy task and performing it on each store actions through Interceptors could be very heavy. Scheduling indexing or reindexing for the nigth could be a good idea if your system is not used worldwide and does not neeed absolutly up to date indexes. Have you seen what is going on with slide and DASL ? If not have a look at org.apache.slide.search package JP > If messages of this sort are more appropriate on the dev group, please tell > me so that I can do the right thing next time > Thanks > Akil > [EMAIL PROTECTED] > > > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> > > > -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
