Urls and unicode tags

[adrien-delhorme]

Description

URL for tags list view uses slug as path converter, which only accepts ASCII characters.:

djangocms-blog/djangocms_blog/urls.py

Line 37 in 30dbb72

path("tag/<slug:tag>/", TaggedListView.as_view(), name="posts-tagged"),

But django-taggit allows unicode in tags' slugs.

When a reverse() is done on this URL, it throws an exception:

Exception Type: NoReverseMatch
Exception Value: Reverse for 'posts-tagged' with keyword arguments '{'tag': 'en-été'}' not found. 1 pattern(s) tried: ['fr/blog/tag/(?P<tag>[-a-zA-Z0-9_]+)/\\Z']

Steps to reproduce

1. Create a tag with a label containing unicode chars (for example : "En été")
2. Assign this tag to a Post
3. Add the BlogTagsPlugin on a page

Versions

Python == 3.11
Django == 3.2.25
djangocms-blog == 2.0.7

[adrien-delhorme]

A workaround is to set TAGGIT_STRIP_UNICODE_WHEN_SLUGIFYING = True in your settings.

Tags created via a Post will have a slug without unicode characters. But tags created via Taggit section of django admin will have their slug generated by prepopulated_fields and therefore unicode characters in it... (but it is a django-taggit issue).

[adrien-delhorme]

And to fix tag creation via django's admin, we can use a pre_save signal to remove unicode chars from slugs before saving a Tag instance.

from django.utils.text import slugify
from django.db.models.signals import pre_save

class YourAppConfig(AppConfig):
    ...

    def ready(self):
        from taggit.models import Tag
        pre_save.connect(slug_without_unicode, sender=Tag)

def slug_without_unicode(sender, instance, **kwargs):
  if instance.slug is not None:
      instance.slug = slugify(instance.slug, allow_unicode=False)