HI, Sorry for crossposting here and on StackOverflow but nobody could help me out so far, so I am desperate for a bit more attention.
I am using AppEngine with Restlet to serve my data to a GWT app in the browser as client. The Restlet-GWT edition uses GWT-RPC serialization as the transport format underneath. GWT-RPC serialization relies on the shared source between client and server to serialize/deserialize. Now, after adding a new property to one of the shared source classes, de-serialization started failing. The AppEngine server processed the request correctly with response HTTP 200 / OK and was sending out a correctly serialized object. The client choked each time. After awhile I figured out that the browser was trying to deserialize a cached copy of my object (without the newly added property) and so de-serialization in the browswer failed. Now the question : why is the browser using a cached copy if the server is being hit anyhow ? IMO, this defeats the purpose/advantages of caching, the server and network resources are being consumed and the fresh result is not used ? In case the browswer decides to use a cached copy, I would expect no round-trip to the server. The original question is here : http://stackoverflow.com/questions/10976999/after-hitting-the-server-browser-still-uses-a-cached-version Thx ! Koen ------------------------------------------------------ http://restlet.tigris.org/ds/viewMessage.do?dsForumId=4447&dsMessageId=2970653