On Fri, Jul 4, 2008 at 9:53 AM, Dominik Süß <[EMAIL PROTECTED]> wrote: > ...After thinking a bit I wondered if it wouldn't make sense to define > interceptionpoints for pre and post processing of any servlet in the request > cycle. This would give the developers a consistent and flexible way to > extend the given functionality by their own needs without being to > intrusive....
IIUC this is pretty much what Carsten suggests - my alternative proposal is to let the script take complete control and use only bits of the standard processing, as needed. Both approaches are probably ok, we'd need more concrete use cases to decide I guess (and this is the ASF - whoever does it gets to decide, at least for an initial implementation). -Bertrand > > Regards, > Dominik > > On Thu, Jul 3, 2008 at 4:51 PM, Bertrand Delacretaz <[EMAIL PROTECTED]> > wrote: > >> On Thu, Jul 3, 2008 at 4:45 PM, Dominik Süß <[EMAIL PROTECTED]> >> wrote: >> >> > ...for preprocessing I would see even more usecases - if you think of >> > serverside backendprocesses like booking a flight you just can process >> the >> > request after submitting, but should do the backendaction before >> persisting >> > the node.... >> >> I agree that pre-processing might be needed as well. >> >> Another way of implementing this would be to let the POST script take >> complete control of the request processing, while making the >> SlingPostServlet operations available to it. >> >> The script could then pick and choose which of these operations it >> wants to execute. That might require some refactoring of the >> SlingPostServlet stuff, but I think it's already fairly modular so >> that shouldn't be too bad. >> >> -Bertrand >> > -- Bertrand Delacretaz http://www.codeconsult.ch
