Server error 500 if no accept_language header is present

[o-zander]

If a bot like the google bots who does not set the accept_language header is visiting the page without a language prefix and i18n is activated, following error is thrown:

NoReverseMatch: Reverse for 'pages-details-by-slug' with arguments '()' and keyword arguments '{'slug': ''}' not found.

Only happens if DEBUG is False.

[ryanbagwell]

I just spent the last 7 hours encountering this same issue. It's especially a problem when a third-party service (i.e. Instagram) tries to call a callback url without the header set.

[ryanbagwell]

I actually just figured this out, although I don't completely understand the problem.

In my case, I had set the available Django LANGUAGES setting to use only one - 'en'. However, the default LANGUAGE_CODE setting is 'en-us'. Accordingly, when requests were made without the language header were made, Django assumed it wanted a response in 'en-us'. In my case, no page actually existed in that language, because my pages only existed in 'en'.

The CMS will offer the option to localize the content for every language specified in LANGUAGES. By default, that's every language that Django supports, and results in an admin with a billion language choices. I didn't want that, which is why I set the LANGUAGES to a single choice, and ended up using the wrong language code.

[JHei]

same issue here, solved it by setting LANGUAGES and LANGUAGE_CODE to matching values.

[mrsarm]

Same issue, my languages variables are:

LANGUAGE_CODE = 'es-ar'
LANGUAGES = (
    ('es', gettext('Spanish')),
    ('en', gettext('English')),
)

Like @ryanbagwell says, looks like Django are searching a erroneous language, if the page have both language versions, no problem. Also only occurs when the user is sign-in.

Tested with the latest 3.0.0.beta3 and Python 2.7.

[mrsarm]

@digi604 , according with the official Django documentation (https://django.readthedocs.org/en/1.6.x/ref/settings.html#language-code), this is correct:

LANGUAGE_CODE = 'en-us'

An also this in the same context (https://django.readthedocs.org/en/1.6.x/ref/settings.html#languages):

LANGUAGES = (
    ('de', _('German')),
    ('en', _('English')),
)

But looking your patch, a configuration like this will be raise a ImproperlyConfigured error (digi604@92b9909#diff-9cbb78756a2eee73ef8b1d351f9d2800R197):

if not settings.LANGUAGE_CODE in dict(settings.LANGUAGES):
    raise ImproperlyConfigured(
                    'LANGUAGE_CODE "%s" must have a matching entry in LANGUAGES' % settings.LANGUAGE_CODE
                )

I'm wrong?

[digi604]

If your LANGUAGE_CODE is 'en-us' and in your LANGUAGES you have: 'en' and 'de' then django actually supports 3 languages: 'en-us', 'en' and 'de'. The problem is that 'en-us' is not the same as 'en' and it is possible that request.LANGUAGE_CODE becomes 'en-us' if you don't have the language accept headers set. We have 2 possible solutions: Either we add LANGUAGE_CODE to CMS_LANGUAGES as well... and then you can create pages in 'en-us' and 'en' or we enforce that LANGUAGE_CODE is in LANGUAGES. To be honest i think the django docs in this regard should be updated as most people will not understand what is going on in the back and only encounter 500 after they go live and they receive a request without a accept language header and have a new language to deal with for the first time.

[digi604]

PS. over the years i had soooo many bug reports because of this behavior that i finally decided to enforce the LANGUAGE_CODE in LANGUAGES as it makes absolutely sense in a i18n environment.

[miguelfg]

Hi, I just got this nightmare error, but I am still confused how settings.py should be configured. Mine is as follows:

LANGUAGE_CODE = 'es-ES'
from django.utils.translation import ugettext_lazy as _
LANGUAGES = (
('es', _('Spanish')),
('en', _('English')),
('de', _('German')),
)

It looks good as long as I access specifying 'Accepted-Language' header either from browser or testing it with curl. BUt wihtout a 'Accepted-Language' header , breaks with 500.

What should I tweak??

[optimum-dulopin]

Hi,
I have a similar problem, with

django-cms==3.0.2
django = 1.5.8
django-cms-search==0.6.3
whoosh < 2.5
django-haystack < 2.0

but only when I run command
./bin/django rebuild_index

I seems that it succeed to index my french pages, but them it makes and error

All documents removed.
/eggs/django_haystack-1.2.7-py2.7.egg/haystack/management/commands/update_index.py:84: UserWarning: 'SearchIndex.get_queryset' is pending deprecation & will be removed in Haystack v2. Please rename them to 'index_queryset'.
warnings.warn("'SearchIndex.get_queryset' is pending deprecation & will be removed in Haystack v2. Please rename them to 'index_queryset'.")

Indexing 534 <django.utils.functional.proxy object at 0x29a0810>.
/eggs/Django-1.5.8-py2.7.egg/django/conf/urls/defaults.py:3: DeprecationWarning: django.conf.urls.defaults is deprecated; use django.conf.urls instead
DeprecationWarning)

NoReverseMatch: Reverse for 'pages-details-by-slug' with arguments '()' and keyword arguments '{'slug': ''}' not found.

in my settings.py

LANGUAGE_CODE = 'fr'
LANGUAGES = (
('fr', _('French')),
)

what is really weird is that is I also add english language in LANGUAGES, eg :

LANGUAGES = (
('fr', _('French')),
('en', _('English')),
)
then I don't have any error any more . But I dont need english ! I was also trying with german instead of english, but then the error appear again !

thanks for help