Mailing-List: contact struts-user-help@jakarta.apache.org; run by ezmlm
Message-ID: <39E38289.1671E1EC@s1.com>
Date: Tue, 10 Oct 2000 16:56:41 -0400
From: "Jason Kitchen" <jason.kitchen@s1.com>
MIME-Version: 1.0
To: struts-user@jakarta.apache.org
Subject: Re: Multi-page form
References: <39E2867E.D3496917@mindspring.com> <39E31FE0.86212666@s1.com>
 <39E335E3.5124F6BD@s1.com> <39E33E54.EAD4E3C2@s1.com>
 <39E37F2E.99E05C96@mindspring.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Sean Moore wrote:
> 
> > Jason Kitchen wrote:
> > I still think you're missing the point - maybe I didn't explain myself clearly.
> > The solution you provide above means that each time you move a field from
> > "step 1 form" to "step 2 form" or vice-versa you need to modify the correct
> > "partialValidate" method. What I would like to do is to be able to move fields
> > to another form with no code impact (other than JSP). I think that this is what
> > the struts designers intended (comments from the struts developers ?) but
> > the validations are preventing the relaization of this goal.
> >
> 
> Thanks Jason - that is exactly what I am trying to do.
> Maybe I am missing it but I have not heard a solution that does not
> involve a framework change??
> 
> But Jason mentions this proposed modification
> 
> > In addition I think that this is easily achievable with a few modifications to the
> > struts framework:
> >
> > - Change the ValidatingActionForm interface - the signature to the validate() method
> >   becomes: public String[] validate(HttpServletRequest request);
> >
> > - In ActionServlet.processValidate() we need to pass the HttpServletRequest object
> >   as a parameter when we call the validate method.
> 
> I guess after having the HttpServletRequest you can go thru all of the
> arguments that were submitted and then select the ones to validate.
> This is kind of what struts is already doing for you
> by setting all of the variables in the bean.
> I guess I was hoping for something a little more cleaner.  I
> Just not sure what?
> 
> Am I misunderstanding the solution the modification will allow
> 

I was working through how I could hack it to work. I'm not sure how far along
Craig is with his solution but I agree that something cleaner would be 
appropriate along the same lines as the attribute setting in the current
implementation.

Craig - are you far enough along to be able to elaborate a little more what you
had in mind ?

-- Jason