| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
|
|
|
|
|
|
|
| |
relations.
Moved data loss check when assigning to a reverse one-to-one relation on
an unsaved instance to Model.save(). This is exactly the same change as
e4b813c but for reverse relations.
|
|
|
|
| |
enclosure feed tests
|
|
|
|
|
|
|
|
| |
urlparse() fails with an AttributeError ("'__proxy__' object has no
attribute 'decode'") if reverse_lazy is used to look up the URL
(this is exactly the same problem that caused ticket #18776). The
solution is to use force_str() on the path before handing it to
urlparse().
|
|
|
|
|
|
| |
squashmigrations
Thanks Tim Graham for the review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The change partly goes back to the old behavior for forwards migrations
which should reduce the amount of memory consumption (#24745). However,
by the way the current state computation is done (there is no
`state_backwards` on a migration class) this change cannot be applied to
backwards migrations. Hence rolling back migrations still requires the
precomputation and storage of the intermediate migration states.
This improvement also implies that Django does not handle mixed
migration plans anymore. Mixed plans consist of a list of migrations
where some are being applied and others are being unapplied.
Thanks Andrew Godwin, Josh Smeaton and Tim Graham for the review as well
as everybody involved on the ticket that kept me looking into the issue.
|
|
|
|
| |
list/tuple of decorators.
|
|
|
|
| |
password_reset() view.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Atom feeds.
The ``item_enclosures`` hook returns a list of ``Enclosure`` objects which is
then used by the feed builder. If the feed is a RSS feed, an exception is
raised as RSS feeds don't allow multiple enclosures per feed item.
The ``item_enclosures`` hook defaults to an empty list or, if the
``item_enclosure_url`` hook is defined, to a list with a single ``Enclosure``
built from the ``item_enclosure_url``, ``item_enclosure_length``, and
``item_enclosure_mime_type`` hooks.
|
|
|
|
| |
values.
|
|
|
|
| |
places and max digits.
|
|
|
|
|
|
|
| |
annotations.
This doesn't work in Django 1.8 but was fixed in Django 1.9 as a
side effect of another change.
|
|
|
|
| |
error to a system check.
|
|
|
|
|
|
|
|
| |
queryset chaining fix.
The issue was fixed on master due to the removal of ValuesQuerySet.
Partial forwardport of 7d471fe6626de2b36b97ac04086d8ea307be8288 from stable/1.8.x
|
| |
|
|
|
|
| |
requirements.
|
|
|
|
| |
requirements.
|
|
|
|
| |
on gis backends.
|
| |
|
|
|
|
| |
for timezones on the negative side of UTC.
|
|
|
|
|
|
|
| |
CSRF_COOKIE_DOMAIN.
Thanks Seth Gottlieb for help with the documentation and
Carl Meyer and Joshua Kehn for reviews.
|
| |
|
|
|
|
|
|
|
| |
TemplateSyntaxError strings.
This makes it much easier to diagnose a test failure when all
you have is the stack trace from an uncaught TemplateSyntaxError.
|
|
|
|
|
|
| |
It didn't work anymore since the commit that enabled test
parallelization by default because parallel was equal to 0
at that point.
|
|
|
|
|
|
| |
Since --parallel is documented not to work on Windows, it's better to
ignore it and run without parallelization than to crash. For example
this could simplify cross-platform test scripts.
|
|
|
|
| |
COUNT(*) instead of COUNT('*').
|
|
|
|
| |
field with unhashable default.
|
| |
|
|
|
|
| |
disabled migrations
|
|
|
|
| |
app basis
|
|
|
|
| |
flatten a Context.
|
| |
|
|
|
|
|
|
|
| |
instead of classes.
This allows dynamically-generated attributes to be specified in
checked ModelAdmin attributes without triggering errors.
|
|
|
|
| |
left open
|
|
|
|
|
| |
Düsseldorf now appears as Dusseldorf in the latest version
of the GeoIP2 database.
|
|
|
|
| |
Windows.
|
| |
|
|
|
|
|
|
| |
This new technique is more straightforward and compatible with test
parallelization, where the effective database connection settings no
longer match settings.DATABASES.
|
|
|
|
|
| |
Considering the APIs exercised by these test cases, it's hard to make
them independent.
|
|
|
|
|
|
|
|
| |
parallel.
This requires that each test never alters files in static directories
collected by other tests. The alternative is to add a temporary
directory to STATICFILES_DIRS or a new app to INSTALLED_APPS.
|
|
|
|
| |
This is required for running tests in parallel.
|
|
|
|
|
|
|
| |
connections too heavily.
The previous implementation would result in tests hitting the wrong
database when running tests in parallel on multiple databases.
|
| |
|
|
|
|
|
|
|
|
|
| |
Previously, a traceback would be displayed on exit because:
- using some multiprocessing features creates a temporary directory
- this directory would be inside TMPDIR
- multiprocessing would attempt to remove it when a particular object
was deallocated, after runtests.py had already removed it along with
everything else in TMPDIR.
|
| |
|
| |
|
|
|
|
|
|
|
| |
fixtures test.
This prevented the temporary directory from being removed
on Windows.
|
| |
|
|
|
|
|
|
| |
attribute.
Thanks Jaap Roes for the idea and initial patch.
|
|
|
|
| |
model operations.
|