[thelist] Back button issue with QueryString

Chris W. Parker cparker at swatgear.com
Thu Jan 23 12:48:01 CST 2003


Joshua Olson <mailto:joshua at waetech.com> wrote:

> ----- Original Message -----
> From: "Chris W. Parker" <cparker at swatgear.com>
> Sent: Thursday, January 23, 2003 11:55 AM


> And if they press the back button, what happens?  Let's work through
> this for a second:
>
> 1.  page1 is in their browser history as the "current" page.
> 2.  they transfer to page2 for processing.  page1 is one page back in
> history, but doesn't have the error.
> 3.  ok, so we think we set a cookie
> 4.  depending on how you "send them back to page1" the cookie may or
> may not be set.  If you do a server side relocation (such as with
> cflocation) then the cookie never gets set on the client machine.
> 5.  ... now the process is broken.  Hmmm...

well i'm going from my experience with asp. if i set a cookie and then
Response.Redirect("page1.asp?error=this+error") iirc it will work as
expected. that is, the cookie will be set even though the server
relocates. but then again this is what i remember happening and can't
say that for sure it will happen.

> 1.  The page will be pulled out of the cache and the rendered error
> message will display.  If this is the case, the only way to avoid
> displaying the error message is to use something on the client that
> executes on page display that would somehow remove the error message.
> JS is one such technology that could do this.
>
> 2.  The page will be rerequested from the server.  But, since the URL
> is in tact, it'll STILL come back with the error message.  The only
> way to do this would be to somehow let the server know it needs to
> ignore the error message request.  This could be done with normal
> cookies/session variables.  Disable cookies or session variables and
> this wouldn't work.
>
> If you cannot dictate that scenario #2 happens, then you are
> essentially stuck using client side scripting of some sort if you
> intend to keep things somewhat simple.

hmmm... i see what you are saying.

> I'm afraid that your suggestion may not solve this problem yet.

yes.
thanks for the explanation.

chris.



More information about the thelist mailing list