Vadim Gritsenko pisze: > On Dec 19, 2007, at 2:00 PM, Grzegorz Kossakowski wrote: > >> The check has been introduced in order to handle polymorphism of >> servlet services > > Shouldn't it explicitly check for 404 only? And, for Cocoon servlet, for > unhandled ResourceNotFoundException? > > Any of 1XX-3XX, 5XX are clearly not an indication for calling super, > IIUC. Seems that only 404 ('The server has not found anything matching > the Request-URI.') should be. > > Even then, if 'super' is not available, it should simply return that 404 > without messing with it... Right? There should be no > wrappedResponse.reset() is super is null. There also should be no "throw > null;", if there were no ServletException.
Sorry for not responding earlier, I must have missed your e-mail somehow. Yep, Vadim you are right on all of your points. The whole snippet is broken (I'm not sure why I committed the patch without noticing it) and as I already pointed out in JIRA we will probably need to buffer whole response. Also, in r611596 I added checks that can easily exhibit bad behaviour in an isolated environment. Anyone willing to take care of fixing it? I'm quite busy ATM... -- Grzegorz Kossakowski Committer and PMC Member of Apache Cocoon http://reflectingonthevicissitudes.wordpress.com/