On Sun, Aug 5, 2012 at 5:17 PM, Paul Benedict <[email protected]> wrote:

> Philip,
>
> On Sun, Aug 5, 2012 at 1:42 AM, Philip Luppens <[email protected]>
> wrote:
> > On Sun, Aug 5, 2012 at 7:34 AM, Paul Benedict <[email protected]>
> wrote:
> >
> >> I don't think this feature exists but it does in Tapestry (and
> >> probably .NET). It should be possible to serialize the ValueStack on a
> >> redirect, pass it as a 64-bit encoded key, and deserialize it. This
> >> would facilitate greater Redirect-After-Post patterns that require
> >> data points to still be active.
> >>
> >> Thoughts?
> >>
> >
> > I find it an intriguing idea, but wouldn't it imply a serialisation of
> the
> > entire object graph that is attached to your root object and hence a
> > security risk by exposing certain variables?
>
> Yes, you would have to serialize what's in the ValueStack. This would
> be a given. However, a decent programmer should know what he is about
> to do by enabling this kind of feature -- probably via a new
> interceptor. With regards to security, the contents definitely require
> encryption. Base64 encoding isn't stealth.


> > Isn't this something that is already possible with the 'flash' scope? Or
> > does this approach offer certain benefits?
> >
> > Cheers,
> >
> > Phil
>
> Yes, it is a type of flash scope. The only difference is that it's not
> dependent on a user's session.
> https://issues.apache.org/jira/browse/WW-2635


Just a quick remark: wouldn't a specialised object be a better approach for
this, rather than the entire ValueStack? It seems to me this can get pretty
huge and therefore impact performance/scaling more than the session-based
approach. Having a custom object serialised allows for better control, and
I doubt you'd need everything kept in the ValueStack on your
Redirect-After-Post (I might be wrong about that one, but we've been either
storing some minimal info in the flash scope, or adding it as a request
parameter, and it seemed to be 'good enough').

My 2 cents,

Phil


>
>
> Paul
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>


-- 
"We cannot change the cards we are dealt, just how we play the hand." -
Randy Pausch

Reply via email to