I have had to write a parrallel implementation of DigesterDescriptorReader as the JR OCM impl doesn't let me get at the Digester. It also means I have to have a parrallel implementation of DigesterMapperImpl.
It would be good to refactor the JCR OCM code a bit to allow either 1) custom DigesterDescriptorReaders. The code in DigesterDescriptorReader#loadClassDescriptors needs to be split a bit to call out to a hook method for the digester which can be over-ridden or which in turn hooks out to get an EntityResolver if necessary. 2) a EntityResolver e.g. CatalogResolver, to be added to the DigesterMapperImpl and therefore the delegated DigesterDescriptorReader which would then need to attach it to the digester upon instantiation The parrallel impl works great ... I can cache the DTD files locally and just let the CatalogManager.properties stored in the jar direct the Resolver on where to find the Jackrabbit DTDs within the jar itself. -- View this message in context: http://www.nabble.com/OCM-DTD-resolution-tp21833477p21855395.html Sent from the Jackrabbit - Dev mailing list archive at Nabble.com.
