This change:
* Makes the InclusionNode cache-safe by removing render-time side effects
to its nodelist.
* Ensures the render_context stack is properly scoped and reset by updating
the render call to use Template.render rather than Nodelist.render.
Backport of 0808ccce38 from master
Rendering a Jinja template with self in the context threw an error.
While self is a reserved variable in Jinja, including self in the
context is not an error, so Django should respect that.
Backport of 4ea1909d3c from master
Calling Migration.mutate_state() now also allows to do in_place
mutations in case an intermediate state is thrown away later.
Thanks Anssi Kääriäinen for the idea, Ryan Hall for parts of the patch,
and Claude Paroz and Tim Graham for the review
Backport of 57dc8dd3fa from master
This may cause some backwards compatibility issues, but may also
resolve security issues in third party projects that fail to heed warnings
in our documentation.
Thanks Markus Holtermann for help with tests and docs.
Backport of fa350e2f30 from master
The query used a construct of qs.annotate().values().aggregate() where
the first annotate used an F-object reference and the values() and
aggregate() calls referenced that F-object.
Also made sure the inner query's select clause is as simple as possible,
and made sure .values().distinct().aggreate() works correctly.
Backport of fb146193c4 from master
These cached properies were causing problems with pickling, and in
addition they were confusingly defined: field.rel.model._meta was
not the same as field.rel.opts.
Instead users should use field.rel.related_model._meta inplace of
field.rel.opts, and field.rel.to._meta in place of field.rel.to_opts.
Backport of f95122e541 from master
Switched from an adjancency list and uncached, iterative depth-first
search to a Node-based design with direct parent/child links and a
cached, recursive depth-first search. With this change, calculating
a migration plan for a large graph takes several seconds instead of
several hours.
Marked test `migrations.test_graph.GraphTests.test_dfs` as an expected
failure due to reaching the maximum recursion depth.
Backport of 78d43a5e10 from master
The new signature enables better support for routing RunPython and
RunSQL operations, especially w.r.t. reusable and third-party apps.
This commit also takes advantage of the deprecation cycle for the old
signature to remove the backward incompatibility introduced in #22583;
RunPython and RunSQL won't call allow_migrate() when when the router
has the old signature.
Thanks Aymeric Augustin and Tim Graham for helping shape up the patch.
Refs 22583.
Conflicts:
django/db/utils.py
Backport of bed504d70b from master
As suggested by Anssi. This has the slightly strange side effect of
passing the expression to Expression.convert_value has the expression
passed back to it, but it allows more complex patterns of expressions.
Backport of 32d4db66b9 from master
Joint effort between myself, Josh, Anssi and Shai.
Conflicts:
django/db/models/query.py
tests/model_fields/models.py
Backport of 4755f8fc25 from master.
This opens more possibilities, like accessing context.template.origin.
It also follows the chain of objects instead of following a shortcut.
Backport of 1bfcc95 from master
Workaround for http://bugs.python.org/issue20747.
In some corner cases, Python 2 inserts a newline in a header value
despite `maxlinelen` passed in Header constructor.
Thanks Tim Graham for the review.
Backport of efb1f99f94 from master.
Swapped out models don't have a _default_manager unless they have
explicitly defined managers. ModelState.from_model() now accounts for
this case and uses an empty list for managers if no explicit managers
are defined and a model is swapped out.
Backport of 15dc8d1c9d from master
Instead of naively reloading only directly related models (FK, O2O, M2M
relationship) the project state needs to reload their relations as well
as the model changes as well. Furthermore inheriting models (and super
models) need to be reloaded in order to keep inherited fields in sync.
To prevent endless recursive calls an iterative approach is taken.
Backport of b29f3b5120 from master
Explicitly checking for django.template.Template subclasses is
preferrable to duck-typing because both the django.template.Template and
django.template.backends.django.Template have a render() method.
Thanks spectras for the report.
Backport of 47ee7b48 from master
Previously Django only checked for the table name in CreateModel
operations in initial migrations and faked the migration automatically.
This led to various errors and unexpected behavior. The newly introduced
--fake-initial flag to the migrate command must be passed to get the
same behavior again. With this change Django will bail out in with a
"duplicate relation / table" error instead.
Thanks Carl Meyer and Tim Graham for the documentation update, report
and review.
Backport of f287bec583 from master
This allows using a UUIDField primary key along with the JSON session
serializer.
Thanks to Trac alias jamesbeith for the report and Simon Charette
for the initial patch.
Backport of 0f7f5bc9e7 from master
This is a performance optimization and also fixes test errors with the
upcoming merge of contrib tests into tests/. The tests failed on MySQL
because the models with GeometryField were being checked but the
non-GIS MySQL backend didn't know how to handle them.
Backport of 197dd4b8f1 from master
During the autumn DST change, the template engine would fail to convert
the naive representation of now when USE_TZ = True. Passing now in UTC
eliminates the issue.
Thanks mbertheau and tricoder42.
Refs #23714.
Backport of c7a6996df7 from master
This reduces the length of rope RequestContext gives users to hang
themselves with.
Thanks Alex Hill for the report and Tim Graham for the review.
Backport of f2c104a from master.
This significantly improves performance on PyPy. The previous
implementation would generate a new class on every single request,
which is relatively slow.
Backport of 8099d33b65 from master
This dramatically improves performance on PyPy. The following benchmark:
python -mtimeit -s "from django.utils.functional import allow_lazy; from django.utils.translation import ugettext_lazy; f = allow_lazy(lambda s: s, str)" "f(ugettext_lazy('abc'))"
goes from 390us per loop to 165us.
Backport of 82e0cd1571 from master
This script uses the unmaintained hotshot module (gone on Python 3)
and doesn't seem to be Django specific in any way.
Backport of 388d986b8a from master
If importing or initializing a template backend fails, attempting to
access this template backend again must raise the same exception.
Backport of 44ad6915 from master
In addition to ImproperlyConfigured, Engine.get_default() may also raise
ImportError or other exceptions. It's better to catch all exceptions in
places where the default engine isn't strictly required.
Backport of 27f9ff45 from master
The function no longer flushes zfile after each write as doing so can
lead to the gzipped streamed content being larger than the original
content; each flush adds a 5/6 byte type 0 block. Removing this means
buf.read() may return nothing, so only yield if that has some data.
Testing shows without the flush() the buffer is being flushed every 17k
or so and compresses the same as if it had been done as a whole string.
Backport of caa3562d5b from master
make_bytes() assumed that if the Content-Encoding header is set, then
everything had already been dealt with bytes-wise, but in a streaming
situation this was not necessarily the case.
make_bytes() is only called when necessary when working with a
StreamingHttpResponse iterable, but by that point the middleware has
added the Content-Encoding header and thus make_bytes() tried to call
bytes(value) (and dies). If it had been a normal HttpResponse,
make_bytes() would have been called when the content was set, well
before the middleware set the Content-Encoding header.
This commit removes the special casing when Content-Encoding is set,
allowing unicode strings to be encoded during the iteration before they
are e.g. gzipped. This behaviour was added a long time ago for #4969 and
it doesn't appear to be necessary any more, as everything is correctly
made into bytes at the appropriate places.
Two new tests, to show that supplying non-ASCII characters to a
StreamingHttpResponse works fine normally, and when passed through the
GZip middleware (the latter dies without the change to make_bytes()).
Removes the test with a nonsense Content-Encoding and Unicode input - if
this were to happen, it can still be encoded as bytes fine.
Backport of 250aa7c39b from master.
The method is mainly intended for use with UUIDField. For UUIDField we
want to call the field's default even when primary key value is
explicitly set to None to match the behavior of AutoField.
Thanks to Marc Tamlyn and Tim Graham for review.
Backport of 8adc59038c from master
Claude Paroz
Tim Graham
Tim Heap
Markus Holtermann
Andriy Sokolovskiy
Moritz Sichert
Josh Smeaton
Pakal
Jon Walsh
Karl Hobley
John Giannelos
Marten Kenbeek
Sævar Öfjörð Magnússon
Erik Romijn
Anssi Kääriäinen
Jean-Louis Fuchs
Benjamin Wohlwend
Mathias André
Michael Angeletti
Stanislas Guerra
Emin Mastizada
Michael Manfre
Aymeric Augustin
Aymeric Augustin
Don Kirkby
foresmac
Loic Bistuer
Marc Tamlyn
Alex Vidal
Shai Berger
Adam Chainz
Preston Timmons
Sergey Fedoseev
Nik Nyby
Lukas Klein
Collin Anderson
Daniel Pyrathon
jMyles
Aron Podrigal
Alex Gaynor
Varun Sharma
Matthew Somerville
mlavin
Riccardo Magliocchetti
Simon Charette