On 2/7/06, Simon Belak <[EMAIL PROTECTED]> wrote: > > Kevin Dangoor wrote: > > As you can see from the above history, I've tended to treat validation > > errors and exceptions differently. They seem like different beasts to > > me. Validation errors occur through normal program operation, because > > users don't always enter what you'd expect. Exceptions represent an > > honest-to-goodness problem with the system. > > > > Unless there are realistic use cases where people expect exceptions to > > be raised with the same sort of frequency as validation errors, I'd > > like to make sure that we handle validation errors as the primary case > > and then exception catching as something that can be done via > > error_handler(). Does that make sense? > > I for one prefer a more generalised approach, unifying anything causing > ruptures in control-flow to be handled the same. I have already proposed > to Jeff to make Identity raise a "normal" exception and move the > (redirection) logic to an error handler but haven't heard from him yet. > In my mind, ideally any and all (authentication, input, model, ... > -related) error handling would be separated from actual [controller] > functionality. This has undoubtedly affected my implementation of error > handling, in retrospective I should have taken a more natural POV. I > would like this to be nicely integrated into the whole control-flow but > I certainly do not want to introduce something a majority of > users/developers will find cumbersome or not intuitive.
The generalized error_hander mechanism is fine and flexible on the whole. But, I'd still contend that validation errors are of a different breed than actual programming exceptions and deserve mildly special treatment as a result. I'd only support moving identity's redirect logic to an error handler if you didn't have to explicitly put references to it all over the place. (@identity.require is nice and easy, and the control flow for the identity piece does not involve the controller, so it's still clean...) > Since I am writing a wish list, I might as well mention this. I do not > like redirects. It would feel much more natural to just simply call a > method. If this method happens to be exposed, the URL should be > rewritten behind the scenes. What do you mean by "the URL should be rewritten behind the scenes"? There are actually certain advantages to doing a real browser redirect (users don't keep posting the same form over and over again, for example). Kevin
