Regression in f387d024fc75569d2a4a338bfda76cc2f328f627.
Just like `OrderByList` the `ExpressionList` expression used to wrap
`Window.partition_by` must implement `get_group_by_cols` to ensure the
necessary grouping when mixing window expressions with aggregate
annotations is performed against the partition members and not the
partition expression itself.
This is necessary because while `partition_by` is implemented as
a source expression of `Window` it's actually a fragment of the WINDOW
expression at the SQL level and thus it should result in a group by its
members and not the sum of them.
Thanks ElRoberto538 for the report.
Adjustments made to solve_lookup_type to defer the resolving of
references for summarized aggregates failed to account for similar
requirements for lookup values which can also reference annotations
through Aggregate.filter.
Regression in b181cae2e3697b2e53b5b67ac67e59f3b05a6f0d.
Refs #25307.
Thanks Sergey Nesterenko for the report.
This adds aria-describedby for widgets rendered in a fieldset such as
radios. aria-describedby for these widgets is added to the <fieldset>
element rather than each <input>.
`id_for_label` is blank for widgets with multiple inputs such as radios
and multiple checkboxes. Therefore , `help_text` for fields using these
widgets cannot currently be associated using `aria-describedby`.
`id_for_label` is being used as a guard to avoid incorrectly adding
`aria-describedby` to those widgets.
This change uses `auto_id` as the unique identified for the fields
`help_text`. A guard is added to avoid incorrectly adding
`aria-describedby` to inputs by checking the widget's `use_fieldset`
attribute. Fields rendered in a `<fieldset>` should have
`aria-describedby` added to the `<fieldset>` and not every `<input>`.
Added a note about the potential race condition in prefetch_related()
that could produce an inconsistent result, one that does not correspond
to any point in the database history.
This trims and clearly delineates the how to guide from the subsequent
explanation with additional subheadings.
These changes have been discussed with Daniele Procida
at the DjangoCon US 2023 sprints.
Regression in 59bea9efd2768102fc9d3aedda469502c218e9b7,
complements e5c844d6f2a4ac6ae674d741b5f1fa2a688cedf4.
Refs #28477, #34551.
Thanks Haldun Komsuoglu for the report.
