…_data in storage is empty.

Backport of 97121cb from master.

Backport of 17f8496 from master.

Thanks Jørgen Abrahamsen.

Backport of 0ca2d1e.

…dmin.

Thanks Daniele Procida.

Backport of 4720117 from master

…uest` parameter.

5362134 from master.

1bbd36a from master.

Thanks Tadas Dailyda for the report and the patch.
Backport of 2f8ab2f from master.

Thanks sergzach for the suggestion.

Backport of 43f89e0 from master

Backport of c9b577e from master

Thanks Dima Pravdin for the report.

Backport of 74d72e2.

Backport of 83d0cc5.

Thanks Daniele Procida.

Backport of d406afe from master

Thanks grossmanandy at bfusa com and Simon Charette.

Backport of 1dd7492.

…n INSTALLED_APPS

Thanks dpravdin and Jordan Messina.

Backport of 0375244 from master

Backpatch of d194f29

…n the auth app.

This is needed to support custom user models which don't define a manager
named `objects`.

Backport of cc4de61 from master.

Thanks wim@ for the suggestion.

Backport of 5b2d9ba from master

Thanks epicserve for the suggestion.

Backport of 214fb70 from master

Backport of 0db8627 from master

The regression was caused by using .none() when querying for related
models, and the origin field's value was None. This resulted in missing
custom related manager subclass as .none() returns plain QuerySet.

This isn't backport from master, in master .none() correctly preserves
the queryset's class.

Patch provided by Simon Charette, with some minor polish by committer.

Thanks will@ for the suggestion.

Backport of 0de2645 from master

Backport of 71c8539 from master

QuerySet.none() returns an instance of EmptyQuerySet, which may have
undesirable side effects in the presence of custom query set classes.

The implementation of .none() was refactored in master to have the same
effect as .filter(pk__in=[]).

Refs #19652.

Thanks Simon Charrette for the report.

…best practices.

Thanks charettes.

Backport of eaa716a from master

Thanks foo@ for the report and Florian Apolloner for the review.

Backport of eafc036 from master

…from_argv

Thanks Stefan Koegl for the report and Simon Charette for the review.
Backport of b9c8bbf from master.

5b99d5a from master

Thanks sunsongxp@ for the report.

Backport of 537d44b from master

Backport of ee26797 from master

Thanks Gavin Wahl.

Backport of 23e319d.

… content_type.

Thanks Tim for the report and initial patch.

Backport of 89cb771 from master.

Partial (docs only) backport of 47ddd6a from master.

Thanks rafales.

Backport of 9a4a1ce.

Thanks Aramgutang for the report. Refs django#17228.
Backport of 56e5531 from master.

Refs django#16218.

Backport of a0c67c6.

Thanks Chris Rebert for the report.

Backport of d75a54c.

Backport of c6560e4 from master

1ee40f2 from master.

5f7eecd from master.

…unsalted MD5 passwords

Thanks apreobrazhensky at gmail.com for the report.
Backport of 63d6a50 from master.

Backport of 1f8e7bb from master.

This test was randomly failing depending on the library environment.
Backport of a1c470a from master.

Thanks Tiberiu Ana for the report and patch.

Backport of fdaaa24 from master

…lAdmin.get_form() method is overridden.

Backport of 0694d21

… folder. Thanks to camilonova, Russell Keith-Magee, Aymeric Augustin and Ramiro Morales for the feedback.

Backport of c9c40bc

Backport of 2c173ff from master

Thanks Michisu for the report.
Backport of 55c585f from master.

Also changed formating of `DATE_INPUT_FORMATS` and
`TIME_INPUT_FORMATS` for readability.

Thanks minddust for the report!

Backport of 5c70299 from master.

…upersuser

Thanks Baptiste Mispelon for the report and Preston Holmes for the review.
Backport of 2390fe3 from master.

Thanks Aymeric for the suggestion.

Backport of afa3e16 from master

The new regex seems not stable enough for being released. Stripping
with regex might need reevaluation for the next release.
Refs #19237.

… failures with pypy.

Backport of 6afc85a

…links and markup.

Backport of 43efefa from master

Thanks Daniele Procida.

Backport of aa85ccf from master.

…ts app.

Backport of 64623a2.

…name fields

Thanks to Aymeric and Carl for the review.

line was present on master from 1520748

…lbackMiddleware

Thanks Aymeric Augustin for the report and directions for the patch.
Backport of 2ed90ea from master.

… causing spurious failures with Pypy and Postgres.

Backport of db09a2d

…out/err

Thanks xian at mintchaos.com for the report.
Backport of 0201b9d from master

…rsionadded/changed directives

Thanks Aymeric for the patch.

Backport of d93edff from master.

2b5f848 based its changes on #19061
that made the is_active attribute mandatory for user models.
The try/except was not removed for the groups_for_user function.

refs #19780

Backpatch of f556df9. Backpatching
these tests so that it will be easier to backpatch the fix for #19707.

Backpatch of a4e97cf

PostgreSQL and Oracle always support transactions, so the create table
and rollback test is non-necessary on those DBs.

Thanks to shai for report.

Backpatch of 0478780

When a query had a complex where condition (a condition targeting more
than the base table) a subquery was used for deletion. However, the
query had default ordering from the model's meta and Oracle doesn't
work with ordered subqueries.

The regression was caused by fast-path deletion code introduced in
1cd6e04 for fixing #18676.

Thanks to Dylan Klomparens for the report.

Backpatch of 8ef3235

Backpatch of 604d876.

Refs django#15295.
Backport of 8fbc20b from master.

The change in f105fbe made through=None
m2m fields fail in cases where they worked before. It isn't possible to
create such fields using public APIs. The fix is trivial, so it seems
worth adding it.

This is not a backport from master. Master has gotten enough other
changes to related fields internal API that this fix alone isn't enough
to do any good.

The reason was that the except clause needed to remove a connection
from the django.db.connections dict, but other parts of Django do not
expect this to happen. In addition the except clause was silently
swallowing the exception messages.

Refs #19707, special thanks to Carl Meyer for pointing out that this
approach should be taken.

…trings.

This fixes two tests in admin_views which were failing on Python 3, but only if
the tests were run with docutils installed.

Backport of 668d0b8 from master.

Thanks chrismedrela for the report.

Refs django#6262, 44b9076 and 4d94c0c.

826d9de from master.

112c6e9 from master.

Thanks Anssi and Marc Tamlyn for reviewing.

Fixes #19803.

0560bfb from master.

Thanks Marek Brzóska for the report.

f179a51 from master.

…IELD on custom User models.

Thanks to Claude Peroz for the draft patch.

(cherry picked from commit f5e4a69)

…ates.

Backport of 7d5e35c

Python 2 getpass on Windows doesn't accept unicode, even when
containing only ascii chars.
Thanks Semmel for the report and tests.
Backport of 02e5909 from master.

Fixes failure introduced in 02e5909.

Points in the test fixtures have 20 as max coordinate.
Backport of 87854b0 from master.

…n integer primary key.

Thanks to Kaloian Minkov for the reminder about this undocumented requirement.

(cherry picked from commit 91c26ea)

Thanks Danilo Bargen for the report.
Backport of 6323616 from master.

…mentation

Backport of 976dc07 from master

Thanks diegueus9 for the report and itsallvoodoo for the draft patch.

Backport of 7a80904 from master

…y_key from IntegerField to AutoField.

Thanks Keryn Knight.

Backport of 218bbef from master

Backport of 9c2066d from master.

Thanks julien.aubert.mail@ for the report and James Pic for the patch.

Backport of 64d0f89 from master

…er validation.

This is a security fix; disclosure and advisory coming shortly.