Dale Newfield <d...@newfield.org> wrote: > Greg Lindholm wrote: > > Not sure why you think I don't understand referer. > > I already quoted the bits that illustrate this perceived misunderstanding, > but since you apparently want me to be more specific: >
Actually I just wanted to move off the whole referer topic since It was not likely going to be of any help to the OP. > > > Greg Lindholm wrote: >>> >>>> The referer header will contain the URL of the last request >>>> >>> > It won't contain the url of the last request. The last request > (chronologically, since you don't specify any other binding for 'last') was > likely an image or style sheet or something to fill in something on the > previously rendered page, or even something in another tab. > Yes I gave an imprecise usage of "last request". Was it really confusing to you? Did you really think I meant the last chronological request received by the server? > > > which will be the action URL not the jsp page that rendered the >>>> result. >>>> >>> By "action URL" I simply meant the URL contains the name of the "action" and not the name of the jsp that rendered the page. Sorry if this was confusing. > > This is completely dependent upon how you handle posts. If you use the > post-redirect style, it won't be the "action url" (I'm not sure what you > mean by that phrase, but I assume you're differentiating those urls from > ones that are not side-effecting), but the GET url that resulted in the > display of the form. Assuming post-redirect-only-if-successful, then it'll > be the url of the step before this one: "submit-multi-page-form-step-2" when > the step that's failing validation is "submit-multi-page-form-step-3" (or > "submit-multi-page-form-3" if this step failed validation again, which > illustrates where this path will break down). > Yes, very nice description, so what you are saying is the referer header isn't going to be of any use in solving the OP problem? > > > So this wouldn't help the OP if I understood his problem. >>>> >>> > The problem is that the OP doesn't understand his problem. If the > validation step knows what the result page should be, he shouldn't be > returning "input", but rather the name of the logical step to be rendered. > He's right that the code shouldn't need to know what jsp page renders that > step, just the logical name of the page, fleshed out in the .xml file. > > -Dale > Yes your right but I was trying to be helpful. Can we just stop this OT discussion now?
