Hello, After a day of debugging I came to the finding that at some places the request and response gets faked after a retry occurred. zope.app.testing.functional calls the publisher like this: publish(request, handle_errors=handle_errors) deep in publish a new instance of request is created in case of a retry old one dropped, works on the new one def publish(request, handle_errors=True): ... return request (returns the newly instantiated request) back to zope.app.testing.functional: publish(request, handle_errors=handle_errors) it just drops the return value I found the following likely places that could be affected: zope.app.testing.functional response = ResponseWrapper(request.response, path) if env.has_key('HTTP_COOKIE'): self.loadCookies(env['HTTP_COOKIE']) publish(request, handle_errors=handle_errors) self.saveCookies(response) ... response = ResponseWrapper(request.response, path) publish(request, handle_errors=handle_errors) return response zope.server.ftp.publisher.py: response = request.response publish(request) return response.getResult() zope.app.twisted.ftp.utils.py: response = request.response publish(request) return response.getResult() zope.app.debug.debug.py: request = self._request(path, stdin, *args, **kw) getStatus = getattr(request.response, 'getStatus', lambda: None) _publish(request Anybody for/against fixing those? -- Best regards, Adam Groszer mailto:agroszer@gmail.com -- Quote of the day: This dead of midnight is the noon of thought, And Wisdom mounts her zenith with the stars. - Lady Barbauld