I've found very odd thing about using reverse in Django 1.2.1.
I have:
myapp/
views.py
urls.py
in urls.py
from django.conf.urls.defaults import *
urlpatterns = patterns('myapp.views',
url(r'^$', 'browse'),
)
in views.py
from django.shortcuts import render_to_response
from django.core.urlresolvers import reverse
print reverse('myapp.views.browse') # <----- this will print correct value
def browse (request):
print reverse('myapp.views.browse') # <----- this fails with exception
return render_to_response('myapp/browse.html')
When I put reverse method anywhere outside the view method (browse - in this case) I get an exception in every further use of reverse or {% url %} tag.
NoReverseMatch at /
Reverse for 'myapp.views.browse' with arguments '()'
and keyword arguments '{}' not found.
WTF? When I comment/delete the print line outside browse() , second print line inside browse() magically start working!
The most basic case is:
class MyForm(forms.Form):
field = forms.CharField(default=reverse(....))
def some_view(request):
print reverse(...)
....
1) I define a class in main-scope that is initialized when django initialize (and runs reverse) 2) When a request comes the some_view function has been triggered and it evaluates the reverse function again (and fails with exception).
I don't see anything bad at all in this approach. Why not to initialise some values in the django main-scope with results of the reverse() function ?