| Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
unsupported.
Executing a DDL statement during a transaction on backends that don't support
it silently commits, leaving atomic() in an incoherent state.
While schema_editor.execute() could technically be used to execute DML
statements such usage should be uncommon as these are usually performed through
the ORM. In other cases schema_editor.connection.execute() can be used to
circumvent this check.
Thanks Adam and Tim for the review.
|
|
|
|
|
|
Fixed Python 2 regression in 6dbe56ed7855f34585884a2381fb1cec22ddc824.
Thanks Simon Charette for the implementation idea.
|
|
Thanks Shivang Bharadwaj for the initial patch.
|
|
|
|
|
|
|
|
non-empty first value.
|
|
|
|
|
|
|
|
Thanks Carl Meyer and Tim Graham for contributing to the patch.
|
|
|
|
header.
|
|
|
|
|
|
|
|
|
|
This allows running foreign key data and schema altering operations in the
same migration on PostgreSQL.
Thanks Tim for review.
|
|
|
|
|
|
reloading after an admin actions failure.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The "needs to have a value for field" messages are incorrect and
reference nonexistent fields since the commit in which they were
introduced (refs #18823).
|
|
starttls(), login(), and other connection methods already call
the method as needed.
|
|
|
|
|
|
get_list_display() is overridden.
|
|
|
|
|
|
|
|
|
|
Thanks Adam Chainz for review and guidance.
|
|
|
|
|
|
Thanks Jaap Roes for the initial patch.
|
|
|
|
|
|
|
|
|
|
These values are duplicated in the super().
|