Our WSGIServer rewrapped the socket errors from server_bind into
WSGIServerExceptions, which is used later on to provide nicer
error messages in runserver and used by the liveserver to see if
the port is already in use. But wrapping server_bind isn't enough since
it only binds to the socket, socket.listen (which is called from
server_activate) could also raise "Address already in use".
Instead of overriding server_activate too I chose to just catch socket
errors, which seems to make more sense anyways and should be more robust
against changes in wsgiref.
Backport of 2ca00faa91 from master.
tearDownClass is not called if setUpClass throws an exception, in our case
this means that LiveServerTestCase leaks LiveServerThread sockets if the
test happens to be skipped later on, and AdminSeleniumWebDriverTestCase
doesn't close it's already open browser window. To prevent this leakage
we catch errors where needed and manually call _tearDownClassInternal.
_tearDownClassInternal should be written as defensively as possible since
it is not allowed to make any assumptions on how far setUpClass got.
This patch should fix the sporadic "Address already in use"-errors on jenkins
and also the "This code isn't under transaction management"-error for sqlite
(also just on jenkins).
After discussion with koniiiik, jezdez, kmtracey, tos9, lifeless, nedbat and
voidspace it was decided that this is the safest approach (thanks to everyone
for their comments and help). Manually calling tearDownClass was shut down
cause we don't know how our users override our classes.
This is a private and very specialized API on purpose and should not be used
without a strong reason!
This patch partially reverts the earlier attempts to fix those issues,
namely:
2fa0dd73b1 and
3c5775d36f
Final note: If this patch breaks in a later version of Django, please be
very careful on how you fix it, you might not see test failures locally.
That said, this patch hopefully doesn't produce even more failures.
Backport of 73a610d2a8 from master.
In Django < 1.6, override_settings restores the settings module that was
active when the override_settings call was executed, not when it was
run. This can make a difference when override_settings is applied to a
class, since it's executed when the module is imported, not when the
test case is run.
In addition, if the settings module for tests is stored alongside the
tests themselves, importing the settings module can trigger an import
of the tests. Since the settings module isn't fully imported yet,
class-level override_settings statements may store a reference to an
incorrect settings module. Eventually this will result in a crash during
test teardown because the settings module restored by override_settings
won't the one that was active during test setup.
While Django should prevent this situation in the future by failing
loudly in such dubious import sequences, that change won't be backported
to 1.5 and 1.4. However, these versions received the "allowed hosts"
patch and they're prone to "AttributeError: 'Settings' object has no
attribute '_original_allowed_hosts'". To mitigate this regression, this
commits stuffs _original_allowed_hosts on a random module instead of the
settings module.
This problem shouldn't occur in Django 1.6, see #20290, but this patch
will be forward-ported for extra safety.
Also tweaked backup variable names for consistency.
Forward port of 0261922 from stable/1.5.x.
Conflicts:
django/test/utils.py
Also moved its documentation to the 'advanced' section. It doesn't
belong to the 'overview'. Same for TransactionTestCase.reset_sequences.
When available_apps is set, after a TransactionTestCase, the database
is now totally empty. post_syncdb is fired at the beginning of the next
TransactionTestCase.
Refs #20483.
This can be used to make Django's test suite significantly faster by
reducing the number of models for which content types and permissions
must be created and tables must be flushed in each non-transactional
test.
It's documented for Django contributors and committers but it's branded
as a private API to preserve our freedom to change it in the future.
Most of the credit goes to Anssi. He got the idea and did the research.
Fixed#20483.
It's useful to be able to list all the (flattened) keys of a
ContextList, to help you figure out why the variable that's supposed
to be there is not.
No .values() or .items() added as the definition for those aren't clear.
The patch is Chris Wilson's patch from pull request 1065 with some
modifications by committer.
The ``DocTestRunner`` and ``OutputChecker`` were formerly in
``django.test.testcases``, now they are in ``django.test.simple``. This avoids
triggering the ``django.test._doctest`` deprecation message with any import
from ``django.test``. Since these utility classes are undocumented internal
API, they can be moved without a separate deprecation process.
Also removed the deprecation warnings specific to these classes, as they are
now covered by the module-level warning in ``django.test.simple``.
Thanks Anssi for the report.
Refs #17365.
SuspiciousOperations have been differentiated into subclasses, and
are now logged to a 'django.security.*' logger. SuspiciousOperations
that reach django.core.handlers.base.BaseHandler will now return a 400
instead of a 500.
Thanks to tiwoc for the report, and Carl Meyer and Donald Stufft
for review.
TestCase._fixture_setup disables transactions so,
in case of an error, cleanup is needed to re-enable
transactions, otherwise following TransactionTestCase
would fail.
Thanks to Preston Timmons for the bulk of the work on the patch, especially
updating Django's own test suite to comply with the requirements of the new
runner. Thanks also to Jannis Leidel and Mahdi Yusuf for earlier work on the
patch and the discovery runner.
Refs #11077, #17032, and #18670.
The AssertNumQueriesContext didn't reset connection's use_debug_cursor
in case there was an exception. This resulted in leaking query strings
into connection.queries. Maximum memory use pre-patch was around 700MB,
post-patch it is around 200MB for Django's test suite.
It isn't necessary to disable set_autocommit since its use is prohibited
inside an atomic block. It's still necessary to disable the legacy
transaction management methods for backwards compatibility, until
they're removed.
Since "unless managed" now means "if database-level autocommit",
committing or rolling back doesn't have any effect.
Restored transactional integrity in a few places that relied on
automatically-started transactions with a transitory API.
enter_transaction_management() was nearly always followed by managed().
In three places it wasn't, but they will all be refactored eventually.
The "forced" keyword argument avoids introducing behavior changes until
then.
This is mostly backwards-compatible, except, of course, for managed
itself. There's a minor difference in _enter_transaction_management:
the top self.transaction_state now contains the new 'managed' state
rather than the previous one. Django doesn't access
self.transaction_state in _enter_transaction_management.
Claude Paroz
Florian Apolloner
Aymeric Augustin
Tim Graham
Oliver Beattie
Marc Tamlyn
Tom Christie
Chris Wilson
Carl Meyer
Preston Holmes
Ramiro Morales
Filipa Andrade
Preston Timmons
Gavin Wahl
Loic Bistuer
konarkmodi
Anssi Kääriäinen
Simon Charette
Hiroki Kiyohara
Julien Phalip
Rafal Stozek