Wagtail 5.1.2 release notes

September 25, 2023

What’s new

Bug fixes

  • Avoid use of ignore_conflicts when creating extra permissions for snippets, for SQL Server compatibility (Sage Abdullah)

  • Ensure sequence on wagtailsearchpromotions_query table is correctly set after migrating data (Jake Howard)

  • Change spreadsheet export headings to match listing view column headings (Christer Jensen, Sage Abdullah)

  • Fix numbers, booleans, and None from being exported as strings (Christer Jensen)

  • Restore fallback on full-word search for snippet choosers and generic index views (Matt Westcott)

  • Restore compatibility with pre-7.15 versions of the Elasticsearch Python library, allowing use of Opensearch (Matt Westcott)

  • Fix error when pickling BaseSiteSetting instances (Matt Westcott)

  • For Python 3.13 support - upgrade Willow to v1.6.2, replace imghdr with Willow’s built-in MIME type detection (Jake Howard)

Upgrade considerations

Search within chooser interfaces requires AutocompleteField for full functionality

In Wagtail 4.2, the search bar within snippet chooser interfaces (and custom choosers created via ChooserViewSet) returned results for partial word matches - for example, a search for “wagt” would return results containing “Wagtail” - if this was supported by the search backend in use, and at least one AutocompleteField was present in the model’s search_fields definition. Otherwise, it would fall back to only matching on complete words. In Wagtail 5.0, this fallback behavior was removed, and consequently a model with no AutocompleteFields in place would return no results.

As of Wagtail 5.1.2, the fallback behavior has been restored. Nevertheless, it is strongly recommended that you add AutocompleteField to your models’ search_fields definitions, to ensure that users can receive search results continuously as they type. For example:

from wagtail.search import index
# ... other imports

class MySnippet(index.Indexed, models.Model):
     search_fields = [