This patch also fixes update_error_dict to better handle the use case described
in this ticket, previously the type of the provided container could be lost in
some conditions.
Thanks Russell Keith-Magee for the report and Tim Graham for review.
Backport of eb7df6b8d7 from master
This patch is two-fold; first it ensure that Django does close everything in
request.FILES at the end of the request and secondly the storage system should
no longer close any files during save, it's up to the caller to handle that --
or let Django close the files at the end of the request.
Backport of e2efc8965edf684aaf48621680ef54b84e116576 from master.
Refs #8620.
If we allow any value to remove form fields then we get name clashes
with method names, media classes etc. There was a backwards
incompatibility introduced meaning ModelForm subclasses with declared
fields called media or clean would lose those fields.
Field removal is now only permitted by using the sentinel value None.
The docs have been slightly reworded to refer to removal of fields
rather than shadowing.
Thanks to gcbirzan for the report and initial patch, and several of the
core team for opinions.
Backport of 9fb0f5dddc4cf7f2d294af1bcde2c359cffd90a5 from master
This reverts commit 4fdd51b73240bf9c8d9472fcc45df699f0714755.
See the ticket for concerns with this implementation; it will be revisited.
Backport of aabceadd7d from master
Avoided introducing a new regex-based SQL splitter in the migrations
framework, before we're bound by backwards compatibility.
Adapted this change to the legacy "initial SQL data" feature, even
though it's already deprecated, in order to facilitate the transition
to migrations.
sqlparse becomes mandatory for RunSQL on some databases (all but
PostgreSQL). There's no API to provide a single statement and tell
Django not to attempt splitting. Since we have a more robust splitting
implementation, that seems like a good tradeoff. It's easier to add a
new keyword argument later if necessary than to remove one.
Many people contributed to both tickets, thank you all, and especially
Claude for the review.
Refs #22401.
Backport of 8b5b199 from master
The ticket was originally about two failing tests, which are
fixed by putting their queries in transactions.
Thanks Tim Graham for the report, Aymeric Augustin for the fix,
and Simon Charette, Tim Graham & Loïc Bistuer for review.
Backport of b990df1d63 from master
Previously, saving a model instance to a non-related field (in
particular a FloatField) would silently convert the model to an Integer
(the pk) and save it. This is undesirable behaviour, and likely to cause
confusion so the validatio has been hardened.
Thanks to @PirosB3 for the patch and @jarshwah for the review.
Change database test settings from "TEST_"-prefixed entries in the
database settings dictionary to setting in a dictionary that is itself
an entry "TEST" in the database settings.
Refs #21775
Thanks Josh Smeaton for review.
Also took the opportunity to slightly refactor gettext options
so as to ease customization by subclassing the command.
Thanks Michal Čihař for the report and initial patch.
GenericRelation now supports an optional related_query_name argument.
Setting related_query_name adds a relation from the related object back to
the content type for filtering, ordering and other query operations.
Thanks to Loic Bistuer for spotting a couple of important issues in
his review.
This feature allows the default `TIMEOUT` Cache argument to be set to `None`,
so that cache instances can set a non-expiring key as the default,
instead of using the default value of 5 minutes.
Previously, this was possible only by passing `None` as an argument to
the set() method of objects of type `BaseCache` (and subtypes).
Allows custom behavior for setting initial form data in ModelAdmin.
By default, initial data is set via GET params. The new method allows
this behavior to be overridden.
Thanks egasimus for the suggestion.
