Hy there,
i just read here that posted data can't be sent by a redirect,
and then found that people are generally not happy with using redirects. And my question is why?
My situation
I have a Django app which starts by rendering a page which gives the choice of using Registered or Anonymous usage.
def start(request):
request.session.flush()
return render_to_response('start.html')
If Registered usage is chosen a simple form is rendered to insert the registration data, and it loops to itself until form is valid after which 'auth/' takes place.
def registration(request):
if request.method == "POST":
form = RegistrationForm(request.POST)
if form.is_valid():
return HttpResponseRedirect('auth/')
else:
return render_to_response('registration.html',
{'form':form})
form = RegistrationForm()
return render_to_response('registration.html',
{'form':form})
If anonymous usage is chosen, the registration page is skipped, and is redirected to 'auth/', the same Django view function like in upper code:
return HttpResponseRedirect('auth/')
What i was hoping to achieve with auth is for it to just set the sessions, write the data to a database and then redirect the next page which is identical to Registered and Anonymous users (only the sessions would differ).
def auth(request):
ipdb.set_trace()
if request.method == "POST":
request.session['user_type'] = 'REG'
request.session['email'] = request.POST['email']
request.session['first_name'] = request.POST['first_name']
RegisteredUser.objects.create(first_name = request.POST['first_name'],
email = request.POST['email'],
company = request.POST['company'])
else:
request.session['user_type'] = 'ANONIM'
return HttpResponseRedirect('next_page')
Of course, when the debugger starts, request.method always returns GET.
My reasons
There is a lots of talking about separating the view from the logic, and i don't find it very readable or loosely coupled when i have a Django view function (if I drop the 'auth/', 'next_page' would have to take it's functionality) which has to
Check for request.post
- If there if the reason for it is form validation of the current page, or if it's from the previous page, in which case write that data to a base
Do additional checks to set the sessions for registered or anonymous user
which I would have to do since using separated functions just for logic that redirects instead of rendering doesn't seem as the generally accepted way.
Can someone please shed some light on things?
UPDATE WITH SOLUTION
Thanks to the answers i figured out the pattern on how to do this. The code for registration is now:
def registration(request):
if request.method == "POST":
form = RegistrationForm(request.POST)
if form.is_valid():
request.session['user_type'] = 'REG'
request.session['email'] = request.POST['email']
request.session['first_name'] = request.POST['first_name']
RegisteredUser.objects.create(first_name = request.POST['first_name'],
email = request.POST['email'],
company = request.POST['company'])
return HttpResponseRedirect('param_select/')
else:
return render_to_response('registration.html',
{'form':form},
context_instance = RequestContext(request))
form = RegistrationForm()
return render_to_response('registration.html',
{'form':form},
context_instance = RequestContext(request))
From now on, I'm using the "Loop into yourself with POST data, and then perform the logic" method.
To demystify the name, thats what the new registration(request) does.
First it renders the form to input the data
Than it is recalled once the data is submitted because of link to itself
form action="" method="post"
This is repeated until correct form data isn't submitted after which the logic is called (in this case, writing to a base and setting the sessions).
It ends with a httpRedirect to a different page leaving the registration completely handled by registration(request).
