Django在代码基本库的 tests 目录中带有自己的测试套件。 我们的政策是确保所有测试始终通过。
我们感谢对测试套件的所有贡献！
\nThe Django tests all use the testing infrastructure that ships with Django for\ntesting applications. See 编写并运行测试 for an explanation of\nhow to write new tests.
\nFirst, fork Django on GitHub.
\nSecond, create and activate a virtual environment. If you're not familiar with\nhow to do that, read our contributing tutorial.
\nNext, clone your fork, install some requirements, and run the tests:
\nLinux / macOS
$ git clone https://github.com/YourGitHubName/django.git django-repo\n$ cd django-repo/tests\n$ python -m pip install -e ..\n$ python -m pip install -r requirements/py3.txt\n$ ./runtests.py\nWindows
...\\> git clone https://github.com/YourGitHubName/django.git django-repo\n...\\> cd django-repo\\tests\n...\\> py -m pip install -e ..\n...\\> py -m pip install -r requirements\\py3.txt\n...\\> runtests.py \nInstalling the requirements will likely require some operating system packages\nthat your computer doesn't have installed. You can usually figure out which\npackage to install by doing a Web search for the last line or so of the error\nmessage. Try adding your operating system to the search query if needed.
\nIf you have trouble installing the requirements, you can skip that step. See\n运行所有测试 for details on installing the optional\ntest dependencies. If you don't have an optional dependency installed, the\ntests that require it will be skipped.
\nRunning the tests requires a Django settings module that defines the databases\nto use. To help you get started, Django provides and uses a sample settings\nmodule that uses the SQLite database. See 使用另一个 settings 配置模块 to\nlearn how to use a different settings module to run the tests with a different\ndatabase.
\nHaving problems? See 错误调试 for some common issues.
\ntoxLink to this heading#Tox is a tool for running tests in different\nvirtual environments. Django includes a basic tox.ini that automates some\nchecks that our build server performs on pull requests. To run the unit tests\nand other checks (such as import sorting, the\ndocumentation spelling checker, and\ncode formatting), install and run the tox\ncommand from any place in the Django source tree:
Linux / macOS
$ python -m pip install tox\n$ tox\nWindows
...\\> py -m pip install tox\n...\\> tox\nBy default, tox runs the test suite with the bundled test settings file for\nSQLite, flake8, isort, and the documentation spelling checker. In\naddition to the system dependencies noted elsewhere in this documentation,\nthe command python3 must be on your path and linked to the appropriate\nversion of Python. A list of default environments can be seen as follows:
Linux / macOS
$ tox -l\npy3\nflake8\ndocs\nisort>=5.1.0\nWindows
...\\> tox -l\npy3\nflake8\ndocs\nisort>=5.1.0\nIn addition to the default environments, tox supports running unit tests\nfor other versions of Python and other database backends. Since Django's test\nsuite doesn't bundle a settings file for database backends other than SQLite,\nhowever, you must create and provide your own test settings. For example, to run the tests on Python 3.7\nusing PostgreSQL:
Linux / macOS
$ tox -e py37-postgres -- --settings=my_postgres_settings\nWindows
...\\> tox -e py37-postgres -- --settings=my_postgres_settings\nThis command sets up a Python 3.7 virtual environment, installs Django's\ntest suite dependencies (including those for PostgreSQL), and calls\nruntests.py with the supplied arguments (in this case,\n--settings=my_postgres_settings).
The remainder of this documentation shows commands for running tests without\ntox, however, any option passed to runtests.py can also be passed to\ntox by prefixing the argument list with --, as above.
Tox also respects the DJANGO_SETTINGS_MODULE environment variable, if\nset. For example, the following is equivalent to the command above:
$ DJANGO_SETTINGS_MODULE=my_postgres_settings tox -e py35-postgres\n对于 Windows 用户应使用：
\n...\\> set DJANGO_SETTINGS_MODULE=my_postgres_settings\n...\\> tox -e py35-postgres\nDjango includes a set of JavaScript unit tests for\nfunctions in certain contrib apps. The JavaScript tests aren't run by default\nusing tox because they require Node.js to be installed and aren't\nnecessary for the majority of patches. To run the JavaScript tests using\ntox:
Linux / macOS
$ tox -e javascript\nWindows
...\\> tox -e javascript\nThis command runs npm install to ensure test requirements are up to\ndate and then runs npm test.
django-docker-boxLink to this heading#django-docker-box allows you to run the Django's test suite across all\nsupported databases and python versions. See the django-docker-box project\npage for installation and usage instructions.
\nsettings 配置模块Link to this heading#The included settings module (tests/test_sqlite.py) allows you to run the\ntest suite using SQLite. If you want to run the tests using a different\ndatabase, you'll need to define your own settings file. Some tests, such as\nthose for contrib.postgres, are specific to a particular database backend\nand will be skipped if run with a different backend.
To run the tests with different settings, ensure that the module is on your\nPYTHONPATH and pass the module with --settings.
The DATABASES setting in any test settings module needs to define\ntwo databases:
A default database. This database should use the backend that\nyou want to use for primary testing.
A database with the alias other. The other database is used to test\nthat queries can be directed to different databases. This database should use\nthe same backend as the default, and it must have a different name.
如果您使用的不是SQLite后端，则需要为每个数据库提供其他详细信息：
\nThe USER option needs to specify an existing user account\nfor the database. That user needs permission to execute CREATE DATABASE\nso that the test database can be created.
PASSWORD 配置选项需要为指定的配置:setting:USER 提供密码。
Test databases get their names by prepending test_ to the value of the\nNAME settings for the databases defined in DATABASES.\nThese test databases are deleted when the tests are finished.
You will also need to ensure that your database uses UTF-8 as the default\ncharacter set. If your database server doesn't use UTF-8 as a default charset,\nyou will need to include a value for CHARSET in the\ntest settings dictionary for the applicable database.
Django的整个测试套件需要花一些时间才能运行，并且，例如，如果您刚刚向Django添加了一个想要快速运行而不运行其他所有功能的测试，则运行每个测试可能是多余的。您可以通过在命令行上将测试模块的名称附加到 runtests.py 上来运行单元测试的子集。
For example, if you'd like to run tests only for generic relations and\ninternationalization, type:
\nLinux / macOS
$ ./runtests.py --settings=path.to.settings generic_relations i18n\nWindows
...\\> runtests.py --settings=path.to.settings generic_relations i18n\nHow do you find out the names of individual tests? Look in tests/ — each\ndirectory name there is the name of a test.
If you want to run only a particular class of tests, you can specify a list of\npaths to individual test classes. For example, to run the TranslationTests\nof the i18n module, type:
Linux / macOS
$ ./runtests.py --settings=path.to.settings i18n.tests.TranslationTests\nWindows
...\\> runtests.py --settings=path.to.settings i18n.tests.TranslationTests\nGoing beyond that, you can specify an individual test method like this:
\nLinux / macOS
$ ./runtests.py --settings=path.to.settings i18n.tests.TranslationTests.test_lazy_objects\nWindows
...\\> runtests.py --settings=path.to.settings i18n.tests.TranslationTests.test_lazy_objects\nYou can run tests starting at a specified top-level module with --start-at\noption. For example:
Linux / macOS
$ ./runtests.py --start-at=wsgi\nWindows
...\\> runtests.py --start-at=wsgi\nYou can also run tests starting after a specified top-level module with\n--start-after option. For example:
Linux / macOS
$ ./runtests.py --start-after=wsgi\nWindows
...\\> runtests.py --start-after=wsgi\nNote that the --reverse option doesn't impact on --start-at or\n--start-after options. Moreover these options cannot be used with test\nlabels.
Some tests require Selenium and a Web browser. To run these tests, you must\ninstall the selenium package and run the tests with the\n--selenium=<BROWSERS> option. For example, if you have Firefox and Google\nChrome installed:
Linux / macOS
$ ./runtests.py --selenium=firefox,chrome\nWindows
...\\> runtests.py --selenium=firefox,chrome\nSee the selenium.webdriver package for the list of available browsers.
\nSpecifying --selenium automatically sets --tags=selenium to run only\nthe tests that require selenium.
Some browsers (e.g. Chrome or Firefox) support headless testing, which can be\nfaster and more stable. Add the --headless option to enable this mode.
如果要运行全套测试，则需要安装许多依赖项：
\nargon2-cffi 16.1.0+
asgiref 3.2.10+ (required)
jinja2 2.7+
Pillow 6.2.0+
pytz (required)
memcached, 加上:ref:supported Python binding <memcached>
sqlparse 0.2.2+ (required)
tblib 1.5.0+
You can find these dependencies in pip requirements files inside the\ntests/requirements directory of the Django source tree and install them\nlike so:
Linux / macOS
$ python -m pip install -r tests/requirements/py3.txt\nWindows
...\\> py -m pip install -r tests\\requirements\\py3.txt\n如果你在安装过程中发生了错误，可能是你的系统缺少一个或多个Python依赖包。请参考安装失败的包的文档或者在网上搜索错误的提示信息。
\nYou can also install the database adapter(s) of your choice using\noracle.txt, mysql.txt, or postgres.txt.
如果要测试memcached缓存后端，则还需要定义一个指向您配置的memcached实例的 CACHES 配置。
To run the GeoDjango tests, you will need to setup a spatial database\nand install the Geospatial libraries.
\n这些依赖项中的每一个都是可选的。 如果您缺少其中的任何一个，则将跳过关联的测试。
\nTo run some of the autoreload tests, you'll need to install the Watchman\nservice.
\n鼓励贡献者对测试套件进行覆盖率测试，以确定需要额外测试的区域。在:ref:testing code coverage 1. 中介绍了coverage代码覆盖度工具的安装和使用。
\nCoverage should be run in a single process to obtain accurate statistics. To\nrun coverage on the Django test suite using the standard test settings:
\nLinux / macOS
$ coverage run ./runtests.py --settings=test_sqlite --parallel=1\nWindows
...\\> coverage run runtests.py --settings=test_sqlite --parallel=1\nAfter running coverage, generate the html report by running:
\nLinux / macOS
$ coverage html\nWindows
...\\> coverage html\n运行Django测试的覆盖率时，随附的 .coveragerc 配置文件将 coverage_html 定义为报告的输出目录，并且还排除了与结果无关的几个目录（Django中包含的测试代码或外部代码）。
Tests for contrib apps can be found in the tests/ directory, typically\nunder <app_name>_tests. For example, tests for contrib.auth are located\nin tests/auth_tests.
master branchLink to this heading#Ensure you have the latest point release of a supported Python version, since there are often bugs in earlier versions\nthat may cause the test suite to fail or hang.
\nOn macOS (High Sierra and newer versions), you might see this message\nlogged, after which the tests hang:
\nobjc[42074]: +[__NSPlaceholderDate initialize] may have been in progress in\nanother thread when fork() was called.\nTo avoid this set a OBJC_DISABLE_INITIALIZE_FORK_SAFETY environment\nvariable, for example:
$ OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES ./runtests.py\nOr add export OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES to your shell's\nstartup file (e.g. ~/.profile).
UnicodeEncodeErrorLink to this heading#If the locales package is not installed, some tests will fail with a\nUnicodeEncodeError.
You can resolve this on Debian-based systems, for example, by running:
\n$ apt-get install locales\n$ dpkg-reconfigure locales\nYou can resolve this for macOS systems by configuring your shell's locale:
\n$ export LANG="en_US.UTF-8"\n$ export LC_ALL="en_US.UTF-8"\nRun the locale command to confirm the change. Optionally, add those export\ncommands to your shell's startup file (e.g. ~/.bashrc for Bash) to avoid\nhaving to retype them.
In case a test passes when run in isolation but fails within the whole suite,\nwe have some tools to help analyze the problem.
\nThe --bisect option of runtests.py will run the failing test while\nhalving the test set it is run together with on each iteration, often making\nit possible to identify a small number of tests that may be related to the\nfailure.
For example, suppose that the failing test that works on its own is\nModelTest.test_eq, then using:
Linux / macOS
$ ./runtests.py --bisect basic.tests.ModelTest.test_eq\nWindows
...\\> runtests.py --bisect basic.tests.ModelTest.test_eq\nwill try to determine a test that interferes with the given one. First, the\ntest is run with the first half of the test suite. If a failure occurs, the\nfirst half of the test suite is split in two groups and each group is then run\nwith the specified test. If there is no failure with the first half of the test\nsuite, the second half of the test suite is run with the specified test and\nsplit appropriately as described earlier. The process repeats until the set of\nfailing tests is minimized.
\nThe --pair option runs the given test alongside every other test from the\nsuite, letting you check if another test has side-effects that cause the\nfailure. So:
Linux / macOS
$ ./runtests.py --pair basic.tests.ModelTest.test_eq\nWindows
...\\> runtests.py --pair basic.tests.ModelTest.test_eq\nwill pair test_eq with every test label.
With both --bisect and --pair, if you already suspect which cases\nmight be responsible for the failure, you may limit tests to be cross-analyzed\nby specifying further test labels after\nthe first one:
Linux / macOS
$ ./runtests.py --pair basic.tests.ModelTest.test_eq queries transactions\nWindows
...\\> runtests.py --pair basic.tests.ModelTest.test_eq queries transactions\nYou can also try running any set of tests in reverse using the --reverse\noption in order to verify that executing tests in a different order does not\ncause any trouble:
Linux / macOS
$ ./runtests.py basic --reverse\nWindows
...\\> runtests.py basic --reverse\nIf you wish to examine the SQL being run in failing tests, you can turn on\nSQL logging using the --debug-sql option. If you\ncombine this with --verbosity=2, all SQL queries will be output:
Linux / macOS
$ ./runtests.py basic --debug-sql\nWindows
...\\> runtests.py basic --debug-sql\nBy default tests are run in parallel with one process per core. When the tests\nare run in parallel, however, you'll only see a truncated traceback for any\ntest failures. You can adjust this behavior with the --parallel option:
Linux / macOS
$ ./runtests.py basic --parallel=1\nWindows
...\\> runtests.py basic --parallel=1\nYou can also use the DJANGO_TEST_PROCESSES environment variable for\nthis purpose.
To avoid polluting the global apps registry and prevent\nunnecessary table creation, models defined in a test method should be bound to\na temporary Apps instance:
from django.apps.registry import Apps\nfrom django.db import models\nfrom django.test import SimpleTestCase\n\nclass TestModelDefinition(SimpleTestCase):\n def test_model_definition(self):\n test_apps = Apps(['app_label'])\n\n class TestModel(models.Model):\n class Meta:\n apps = test_apps\n ...\nSince this pattern involves a lot of boilerplate, Django provides the\nisolate_apps() decorator. It's used like this:
from django.db import models\nfrom django.test import SimpleTestCase\nfrom django.test.utils import isolate_apps\n\nclass TestModelDefinition(SimpleTestCase):\n @isolate_apps('app_label')\n def test_model_definition(self):\n class TestModel(models.Model):\n pass\n ...\nThe decorator can also be applied to classes:
\nfrom django.db import models\nfrom django.test import SimpleTestCase\nfrom django.test.utils import isolate_apps\n\n@isolate_apps('app_label')\nclass TestModelDefinition(SimpleTestCase):\n def test_model_definition(self):\n class TestModel(models.Model):\n pass\n ...\nThe temporary Apps instance used to isolate model registration can be\nretrieved as an attribute when used as a class decorator by using the\nattr_name parameter:
from django.db import models\nfrom django.test import SimpleTestCase\nfrom django.test.utils import isolate_apps\n\n@isolate_apps('app_label', attr_name='apps')\nclass TestModelDefinition(SimpleTestCase):\n def test_model_definition(self):\n class TestModel(models.Model):\n pass\n self.assertIs(self.apps.get_model('app_label', 'TestModel'), TestModel)\nOr as an argument on the test method when used as a method decorator by using\nthe kwarg_name parameter:
from django.db import models\nfrom django.test import SimpleTestCase\nfrom django.test.utils import isolate_apps\n\nclass TestModelDefinition(SimpleTestCase):\n @isolate_apps('app_label', kwarg_name='apps')\n def test_model_definition(self, apps):\n class TestModel(models.Model):\n pass\n self.assertIs(apps.get_model('app_label', 'TestModel'), TestModel)\n