Redirect_to doesn't create history

Rails
1

I have a wizard type flow where on each screen clicking the “continue”
button posts data to the server, then I use redirect_to to show the
main wizard page (which dynamically shows partials based on the
current state). The oddity I’m seeing - and maybe it’s just an HTTP
thing - is that when these redirects happen, there is NO entry in the
browser history - not even for the page that was redirected to. So,
if this happens:

 GET /wizard    (on "page 1" here)
 POST /wizard  (form filled in on "page 1" is processed here,

model’s page parameter changed to “2”, then I do a redirect_to
wizard_url)
GET /wizard (on “page 2” here because the model had it’s page
number changed)

What I see in history is the initial entry for wizard and that’s it.
So if a user is on page 2 (or 3 or 4 or whatever) and hits their back
button, they go back to whatever page was displayed before they first
started the wizard.

Is there a special HTTP code I can set using the :status parameter to
force it to put ALL of the GET requests in the history (even the ones
initiated by redirect)?

Thanks for any help.

jc

2

On Sep 28, 2010, at 9:16 PM, jdc wrote:

model’s page parameter changed to “2”, then I do a redirect_to
force it to put ALL of the GET requests in the history (even the ones
initiated by redirect)?

There might be, but I haven’t heard of it. What your describing
actually makes sense. A redirect implies the page is now elsewhere. So
why save it in the history? That is…

GET /some-page (in history)
GET /some-old-url (results in redirect to /some-new-url)
GET /some-new-url (in history)

Why, normally would you want to see that second request in your browser
history?

I don’t know how you have your app setup, but instead of redirecting,
what about sending back an almost empty page containing just enough
javascript or meta-refresh to take you to the next page.

-philip

3

Thanks for the reply. You’re correct - and in your example I’d expect
to see the following in history:

  • /some-page
  • /some-new-url

That’s not exactly what I was describing though. In my case, think of
it as /some-old-url actually redirecting to /some-page instead of /
some-new-url. Since the browser sees a redirect to /some-page - which
is the same url it’s already on - it seems to say “ok, I didn’t go
anywhere” and not change the history at all. So in the history I only
have:

  • /some-page

From a browser standpoint, this makes since because it doesn’t know
that /some-page is dynamically generated based on state saved in a
DB.

This question really arose from me trying to keep a single URL for
every page in the wizard and just display whatever step the user was
actually on. After thinking about the problem for a bit, I decided
that each wizard step is actually a different page and needs it’s own
URL - it didn’t really make sense to do what I was doing. This solved
the problem completely for me since with different URLs the history is
working the way I’d expect - and code got a lot cleaner too!