I decided to wrap the input stream that would do on the fly filtering. However you really need the entire content in memory to do proper regular expressions and content filtering so I only want to read the entire entity at once if I am going to do filtering.
Anyway I thought I knew where I could put the code but it was never executing. It turns out (after looking further into the source) that the entity is not really created until the doDeserialize() method is called (that is correct right?) In doService() I went out to the server and git the response. I then extracted the entity from that response, wrapped it with another entity that will wrap the input stream and then set that entity into the response going back to the original client. I then look and it seems that after the call to doService() and postprocessResponse() a call to deserialize occurs and within there a new Entity is created and populated. I am still learning the overall design and I don't have it all quite yet (as is obvious) My question to you is: Is creating my own DefaultEntityDeserializer the correct approach based on the way the code is designed or are there other intercept points, wrap points that I can use that I have not yet uncovered? thanks Doug -- What profits a man if he gains the whole world yet loses his soul?
