Question

Does anyone know what Heroku changes on deploy? I’ve got a local dev machine setup with Ubuntu 11.10, and my project installed into a venv, all packages managed via pip. The corresponding requirements.txt is in my repo, and is sent to Heroku on push. Both my local machine and Heroku are accessing the same RDS instance on AWS.

Running the Django dev server locally (manage.py runserver) results in a working page. Attempting to run (via my Procfile) the dev server on Heroku causes an error in the template at {% url foobar %} as follows:

Caught TypeError while rendering: 'ParseResult' object is not callable

My requirements.txt contains the following:

Django==1.3.1
MySQL-python==1.2.3
PIL==1.1.7
amqplib==1.0.2
anyjson==0.3.1
boto==2.2.2
celery==2.5.1
distribute==0.6.24
django-celery==2.5.1
django-kombu==0.9.4
django-picklefield==0.2.0
django-piston==0.2.3
gevent==0.13.6
greenlet==0.3.4
gunicorn==0.14.2
httplib2==0.7.4
kombu==2.1.1
python-dateutil==1.5
wsgiref==0.1.2
zencoder==0.4

Anyone have an idea?

Answer 1

I had a similar problem (actually in the urls.py, not in the template rendering) while using django-piston. I managed to fix it by changing

from django.conf.urls.defaults import url

to

from django.conf.urls.defaults import url as django_url

and then changing all the url() directives in urlpatterns to django_url().

This also fixed up the template thing I was seeing in admin pages. I’m not sure if this is the same thing, but might be a fruitful inquiry.