Reinhard Poetz napisał(a): > > Grek, what was your intention with this change? Since the status-code > property stops working, I reverted your change.
My intention was to have http status code set in every case. Before this change, if serializer (or other component) did not set status code it was undefined after pipeline processing ended. That broke caching of servlet: source in the end. Now, with your change it also can be that status code will be undefined if serializer node is not given any status code, right? The best way would be to set the code at the end of processing only if it was not set earlier down the processing. I have no idea how to do it, though :( In short: reverting my change will break caching of servlet: source in some cases. Also, there was the same problem with HTTP_OK vs HTTP_NOT_MODIFIED status codes, see: https://issues.apache.org/jira/browse/COCOON-2009 for more details. -- Grzegorz Kossakowski http://reflectingonthevicissitudes.wordpress.com/
