Remove redundant Case and When.resolve_expression to delegate composite
expression support to BaseExpression.
Thanks Jacob Tyler Walls for the report and test.
Backport of 00c690efbc0b10f67924687f24a7b30397bf47d9 from main.
Remove redundant Func.resolve_expression and adjust CombinedExpression to
delegate source expression resolving to super() to perform checks against
allows_composite_expressions in a single location.
Backport of a76035e925ff4e6d8676c65cb135c74b993b1039 from main.
Rely on ColPairs and TupleIn which support a single column to be specified
to avoid special casing ForwardManyToOne.get_prefetch_querysets().
Thanks Jacob Walls for the report.
Backport of 626d77e52a3f247358514bcf51c761283968099c from main.
The array fields opt-out heuristic failed to account for sized arrays.
Note that we keep relying on db_type as opposed to performing an ArrayField
instance check against the column's field as there could be other
implementations of model fields that use Postgres arrays to store the
optimization must be disabled for all of them.
Refs #35936.
Thanks Claude Paroz for the report and test.
Backport of 22fc151bb86a553d84c62d7effd289356e9b6c6c from main.
Thank you Sarah Boyce, Carlton Gibson, Tim Schilling and Adam Johnson
for reviews.
Co-authored-by: Christophe Henry <contact@c-henry.fr>
Co-authored-by: David Smith <smithdc@gmail.com>
Co-authored-by: Natalia <124304+nessita@users.noreply.github.com>
Co-authored-by: Matthias Kestenholz <mk@feinheit.ch>
There is no need to call validate_ipv46_address() for
ipaddress.IPv4Address()/ipaddress.IPv6Address() instances since this
relies on trying to create these kind objects from strings, so they will
always be valid.
The usage of time.sleep() could result in the update_or_create() thread winning
the race to create the row if the backend takes a while to create a new
connection in the main thread.
Relying on threading.Event ensures that the flow of execution is systematically
yield back and forth between the main thread and the thread in charge of
performing the background update_or_create().
Thanks Saravana Kumar for the report, and Sarah Boyce and Mariusz
Felisiak for the reviews.
Co-authored-by: Natalia <124304+nessita@users.noreply.github.com>
The previous logic was systematically attempting to retrieve last_insert_id
even for models without an AutoField primary key when they had a GeneratedField
on backends that can't return columns from INSERT.
The issue affected MySQL, SQLite < 3.35, and Oracle when the use_returning_into
option was disabled and could result in either crashes when the non-auto
primary key wasn't an IntegerField subclass or silent misassignment of bogus
insert ids (0 or the previous auto primary key insert value) to the first
defined generated field value.
This logic could only be exercised if the composite primary key included an
AutoField but it's not allowed yet (refs #35957).
It was also slightly broken as it expected the AutoField to always be the first
member of returning_fields.
This avoids many awkward checks against NOT_PROVIDED and provides symmetry
with Field.has_default() which is also the reason why it wasn't made a
property.
Thanks to Bhuvnesh Sharma and Adam Johnson for mentoring this Google
Summer of Code 2024 project. Thanks to Sarah Boyce, David Smith, Jacob
Walls and Natalia Bidart for reviews.
