test —- Regression tests package for Python

注解

The test package is meant for internal use by Python only. It isdocumented for the benefit of the core developers of Python. Any use ofthis package outside of Python's standard library is discouraged as codementioned here can change or be removed without notice between releases ofPython.


The test package contains all regression tests for Python as well as themodules test.support and test.regrtest.test.support is used to enhance your tests whiletest.regrtest drives the testing suite.

Each module in the test package whose name starts with test_ is atesting suite for a specific module or feature. All new tests should be writtenusing the unittest or doctest module. Some older tests arewritten using a "traditional" testing style that compares output printed tosys.stdout; this style of test is considered deprecated.

参见

  • Module unittest
  • Writing PyUnit regression tests.

  • doctest —- 文档测试模块

  • Tests embedded in documentation strings.

Writing Unit Tests for the test package

It is preferred that tests that use the unittest module follow a fewguidelines. One is to name the test module by starting it with test and endit with the name of the module being tested. The test methods in the test moduleshould start with test and end with a description of what the method istesting. This is needed so that the methods are recognized by the test driver astest methods. Also, no documentation string for the method should be included. Acomment (such as # Tests function returns only True or False) should be usedto provide documentation for test methods. This is done because documentationstrings get printed out if they exist and thus what test is being run is notstated.

A basic boilerplate is often used:

  1. import unittest
  2. from test import support
  3.  
  4. class MyTestCase1(unittest.TestCase):
  5.  
  6. # Only use setUp() and tearDown() if necessary
  7.  
  8. def setUp(self):
  9. ... code to execute in preparation for tests ...
  10.  
  11. def tearDown(self):
  12. ... code to execute to clean up after tests ...
  13.  
  14. def test_feature_one(self):
  15. # Test feature one.
  16. ... testing code ...
  17.  
  18. def test_feature_two(self):
  19. # Test feature two.
  20. ... testing code ...
  21.  
  22. ... more test methods ...
  23.  
  24. class MyTestCase2(unittest.TestCase):
  25. ... same structure as MyTestCase1 ...
  26.  
  27. ... more test classes ...
  28.  
  29. if __name__ == '__main__':
  30. unittest.main()

This code pattern allows the testing suite to be run by test.regrtest,on its own as a script that supports the unittest CLI, or via thepython -m unittest CLI.

The goal for regression testing is to try to break code. This leads to a fewguidelines to be followed:

  • The testing suite should exercise all classes, functions, and constants. Thisincludes not just the external API that is to be presented to the outsideworld but also "private" code.

  • Whitebox testing (examining the code being tested when the tests are beingwritten) is preferred. Blackbox testing (testing only the published userinterface) is not complete enough to make sure all boundary and edge casesare tested.

  • Make sure all possible values are tested including invalid ones. This makessure that not only all valid values are acceptable but also that impropervalues are handled correctly.

  • Exhaust as many code paths as possible. Test where branching occurs and thustailor input to make sure as many different paths through the code are taken.

  • Add an explicit test for any bugs discovered for the tested code. This willmake sure that the error does not crop up again if the code is changed in thefuture.

  • Make sure to clean up after your tests (such as close and remove all temporaryfiles).

  • If a test is dependent on a specific condition of the operating system thenverify the condition already exists before attempting the test.

  • Import as few modules as possible and do it as soon as possible. Thisminimizes external dependencies of tests and also minimizes possible anomalousbehavior from side-effects of importing a module.

  • Try to maximize code reuse. On occasion, tests will vary by something as smallas what type of input is used. Minimize code duplication by subclassing abasic test class with a class that specifies the input:

  1. class TestFuncAcceptsSequencesMixin:
  2.  
  3. func = mySuperWhammyFunction
  4.  
  5. def test_func(self):
  6. self.func(self.arg)
  7.  
  8. class AcceptLists(TestFuncAcceptsSequencesMixin, unittest.TestCase):
  9. arg = [1, 2, 3]
  10.  
  11. class AcceptStrings(TestFuncAcceptsSequencesMixin, unittest.TestCase):
  12. arg = 'abc'
  13.  
  14. class AcceptTuples(TestFuncAcceptsSequencesMixin, unittest.TestCase):
  15. arg = (1, 2, 3)

When using this pattern, remember that all classes that inherit fromunittest.TestCase are run as tests. The Mixin class in the example abovedoes not have any data and so can't be run by itself, thus it does notinherit from unittest.TestCase.

参见

  • Test Driven Development
  • A book by Kent Beck on writing tests before code.

Running tests using the command-line interface

The test package can be run as a script to drive Python's regressiontest suite, thanks to the -m option: python -m test. Underthe hood, it uses test.regrtest; the call python -mtest.regrtest used in previous Python versions still works. Running thescript by itself automatically starts running all regression tests in thetest package. It does this by finding all modules in the package whosename starts with test_, importing them, and executing the functiontest_main() if present or loading the tests viaunittest.TestLoader.loadTestsFromModule if test_main does not exist. Thenames of tests to execute may also be passed to the script. Specifying a singleregression test (python -m test test_spam) will minimize output andonly print whether the test passed or failed.

Running test directly allows what resources are available fortests to use to be set. You do this by using the -u command-lineoption. Specifying all as the value for the -u option enables allpossible resources: python -m test -uall.If all but one resource is desired (a more common case), acomma-separated list of resources that are not desired may be listed afterall. The command python -m test -uall,-audio,-largefilewill run test with all resources except the audio andlargefile resources. For a list of all resources and more command-lineoptions, run python -m test -h.

Some other ways to execute the regression tests depend on what platform thetests are being executed on. On Unix, you can run make test at thetop-level directory where Python was built. On Windows,executing rt.bat from your PCbuild directory will run allregression tests.

test.support —- Utilities for the Python test suite

The test.support module provides support for Python's regressiontest suite.

注解

test.support is not a public module. It is documented here to helpPython developers write tests. The API of this module is subject to changewithout backwards compatibility concerns between releases.

This module defines the following exceptions:

  • exception test.support.TestFailed
  • Exception to be raised when a test fails. This is deprecated in favor ofunittest-based tests and unittest.TestCase's assertionmethods.
  • exception test.support.ResourceDenied
  • Subclass of unittest.SkipTest. Raised when a resource (such as anetwork connection) is not available. Raised by the requires()function.

The test.support module defines the following constants:

  • test.support.verbose
  • True when verbose output is enabled. Should be checked when moredetailed information is desired about a running test. verbose is set bytest.regrtest.
  • test.support.is_jython
  • True if the running interpreter is Jython.
  • test.support.is_android
  • True if the system is Android.
  • test.support.unix_shell
  • Path for shell if not on Windows; otherwise None.
  • test.support.FS_NONASCII
  • A non-ASCII character encodable by os.fsencode().
  • test.support.TESTFN
  • Set to a name that is safe to use as the name of a temporary file. Anytemporary file that is created should be closed and unlinked (removed).
  • test.support.TESTFN_UNICODE
  • Set to a non-ASCII name for a temporary file.
  • test.support.TESTFN_UNENCODABLE
  • Set to a filename (str type) that should not be able to be encoded by filesystem encoding in strict mode. It may be None if it's not possible togenerate such a filename.
  • test.support.TESTFN_UNDECODABLE
  • Set to a filename (bytes type) that should not be able to be decoded byfile system encoding in strict mode. It may be None if it's notpossible to generate such a filename.
  • test.support.TESTFN_NONASCII
  • Set to a filename containing the FS_NONASCII character.
  • test.support.IPV6_ENABLED
  • Set to True if IPV6 is enabled on this host, False otherwise.
  • test.support.PGO
  • Set when tests can be skipped when they are not useful for PGO.
  • test.support.PIPE_MAX_SIZE
  • A constant that is likely larger than the underlying OS pipe buffer size,to make writes blocking.
  • test.support.SOCK_MAX_SIZE
  • A constant that is likely larger than the underlying OS socket buffer size,to make writes blocking.
  • test.support.TEST_SUPPORT_DIR
  • Set to the top level directory that contains test.support.
  • test.support.TEST_HOME_DIR
  • Set to the top level directory for the test package.
  • test.support.TEST_DATA_DIR
  • Set to the data directory within the test package.
  • test.support.MAX_Py_ssize_t
  • Set to sys.maxsize for big memory tests.
  • test.support.MISSING_C_DOCSTRINGS
  • Return True if running on CPython, not on Windows, and configurationnot set with WITH_DOC_STRINGS.
  • test.support.HAVE_DOCSTRINGS
  • Check for presence of docstrings.
  • test.support.TEST_HTTP_URL
  • Define the URL of a dedicated HTTP server for the network tests.
  • test.support.ALWAYS_EQ
  • Object that is equal to anything. Used to test mixed type comparison.
  • test.support.LARGEST
  • Object that is greater than anything (except itself).Used to test mixed type comparison.
  • test.support.SMALLEST
  • Object that is less than anything (except itself).Used to test mixed type comparison.

The test.support module defines the following functions:

  • test.support.forget(module_name)
  • Remove the module named module_name from sys.modules and delete anybyte-compiled files of the module.
  • test.support.unload(name)
  • Delete name from sys.modules.
  • test.support.unlink(filename)
  • Call os.unlink() on filename. On Windows platforms, this iswrapped with a wait loop that checks for the existence fo the file.
  • test.support.rmdir(filename)
  • Call os.rmdir() on filename. On Windows platforms, this iswrapped with a wait loop that checks for the existence of the file.
  • test.support.rmtree(path)
  • Call shutil.rmtree() on path or call os.lstat() andos.rmdir() to remove a path and its contents. On Windows platforms,this is wrapped with a wait loop that checks for the existence of the files.
  • test.support.makelegacy_pyc(_source)
  • Move a PEP 3147/488 pyc file to its legacy pyc location and return the filesystem path to the legacy pyc file. The source value is the file systempath to the source file. It does not need to exist, however the PEP3147/488 pyc file must exist.
  • test.support.isresource_enabled(_resource)
  • Return True if resource is enabled and available. The list ofavailable resources is only set when test.regrtest is executing thetests.
  • test.support.python_is_optimized()
  • Return True if Python was not built with -O0 or -Og.
  • test.support.with_pymalloc()
  • Return _testcapi.WITH_PYMALLOC.
  • test.support.requires(resource, msg=None)
  • Raise ResourceDenied if resource is not available. msg is theargument to ResourceDenied if it is raised. Always returnsTrue if called by a function whose name is 'main'.Used when tests are executed by test.regrtest.
  • test.support.systemmust_validate_cert(_f)
  • Raise unittest.SkipTest on TLS certification validation failures.
  • test.support.sortdict(dict)
  • Return a repr of dict with keys sorted.
  • test.support.findfile(filename, subdir=None)
  • Return the path to the file named filename. If no match is foundfilename is returned. This does not equal a failure since it could be thepath to the file.

Setting subdir indicates a relative path to use to find the filerather than looking directly in the path directories.

  • test.support.createempty_file(_filename)
  • Create an empty file with filename. If it already exists, truncate it.
  • test.support.fd_count()
  • Count the number of open file descriptors.
  • test.support.setmatch_tests(_patterns)
  • Define match test with regular expression patterns.
  • test.support.rununittest(*classes_)
  • Execute unittest.TestCase subclasses passed to the function. Thefunction scans the classes for methods starting with the prefix test_and executes the tests individually.

It is also legal to pass strings as parameters; these should be keys insys.modules. Each associated module will be scanned byunittest.TestLoader.loadTestsFromModule(). This is usually seen in thefollowing test_main() function:

  1. def test_main():
  2. support.run_unittest(__name__)

This will run all tests defined in the named module.

  • test.support.rundoctest(_module, verbosity=None, optionflags=0)
  • Run doctest.testmod() on the given module. Return(failure_count, test_count).

If verbosity is None, doctest.testmod() is run with verbosityset to verbose. Otherwise, it is run with verbosity set toNone. optionflags is passed as optionflags todoctest.testmod().

  • test.support.setswitchinterval(interval)
  • Set the sys.setswitchinterval() to the given interval. Definesa minimum interval for Android systems to prevent the system from hanging.
  • test.support.checkimpl_detail(**guards_)
  • Use this check to guard CPython's implementation-specific tests or torun them only on the implementations guarded by the arguments:
  1. check_impl_detail() # Only on CPython (default).
  2. check_impl_detail(jython=True) # Only on Jython.
  3. check_impl_detail(cpython=False) # Everywhere except CPython.
  • test.support.checkwarnings(*filters, _quiet=True)
  • A convenience wrapper for warnings.catch_warnings() that makes iteasier to test that a warning was correctly raised. It is approximatelyequivalent to calling warnings.catch_warnings(record=True) withwarnings.simplefilter() set to always and with the option toautomatically validate the results that are recorded.

checkwarnings accepts 2-tuples of the form ("message regexp",WarningCategory) as positional arguments. If one or more _filters areprovided, or if the optional keyword argument quiet is False,it checks to make sure the warnings are as expected: each specified filtermust match at least one of the warnings raised by the enclosed code or thetest fails, and if any warnings are raised that do not match any of thespecified filters the test fails. To disable the first of these checks,set quiet to True.

If no arguments are specified, it defaults to:

  1. check_warnings(("", Warning), quiet=True)

In this case all warnings are caught and no errors are raised.

On entry to the context manager, a WarningRecorder instance isreturned. The underlying warnings list fromcatch_warnings() is available via the recorder object'swarnings attribute. As a convenience, the attributes of the objectrepresenting the most recent warning can also be accessed directly throughthe recorder object (see example below). If no warning has been raised,then any of the attributes that would otherwise be expected on an objectrepresenting a warning will return None.

The recorder object also has a reset() method, which clears thewarnings list.

The context manager is designed to be used like this:

  1. with check_warnings(("assertion is always true", SyntaxWarning),
  2. ("", UserWarning)):
  3. exec('assert(False, "Hey!")')
  4. warnings.warn(UserWarning("Hide me!"))

In this case if either warning was not raised, or some other warning wasraised, check_warnings() would raise an error.

When a test needs to look more deeply into the warnings, rather thanjust checking whether or not they occurred, code like this can be used:

  1. with check_warnings(quiet=True) as w:
  2. warnings.warn("foo")
  3. assert str(w.args[0]) == "foo"
  4. warnings.warn("bar")
  5. assert str(w.args[0]) == "bar"
  6. assert str(w.warnings[0].args[0]) == "foo"
  7. assert str(w.warnings[1].args[0]) == "bar"
  8. w.reset()
  9. assert len(w.warnings) == 0

Here all warnings will be caught, and the test code tests the capturedwarnings directly.

在 3.2 版更改: New optional arguments filters and quiet.

  • test.support.checkno_resource_warning(_testcase)
  • Context manager to check that no ResourceWarning was raised. Youmust remove the object which may emit ResourceWarning before theend of the context manager.
  • test.support.recordoriginal_stdout(_stdout)
  • Store the value from stdout. It is meant to hold the stdout at thetime the regrtest began.
  • test.support.get_original_stdout()
  • Return the original stdout set by record_original_stdout() orsys.stdout if it's not set.
  • test.support.strippython_strerr(_stderr)
  • Strip the stderr of a Python process from potential debug outputemitted by the interpreter. This will typically be run on the result ofsubprocess.Popen.communicate().
  • test.support.args_from_interpreter_flags()
  • Return a list of command line arguments reproducing the current settingsin sys.flags and sys.warnoptions.
  • test.support.optim_args_from_interpreter_flags()
  • Return a list of command line arguments reproducing the currentoptimization settings in sys.flags.
  • test.support.captured_stdin()
  • test.support.captured_stdout()
  • test.support.captured_stderr()
  • A context managers that temporarily replaces the named stream withio.StringIO object.

Example use with output streams:

  1. with captured_stdout() as stdout, captured_stderr() as stderr:
  2. print("hello")
  3. print("error", file=sys.stderr)
  4. assert stdout.getvalue() == "hello\n"
  5. assert stderr.getvalue() == "error\n"

Example use with input stream:

  1. with captured_stdin() as stdin:
  2. stdin.write('hello\n')
  3. stdin.seek(0)
  4. # call test code that consumes from sys.stdin
  5. captured = input()
  6. self.assertEqual(captured, "hello")
  • test.support.tempdir(_path=None, quiet=False)
  • A context manager that creates a temporary directory at path andyields the directory.

If path is None, the temporary directory is created usingtempfile.mkdtemp(). If quiet is False, the context managerraises an exception on error. Otherwise, if path is specified andcannot be created, only a warning is issued.

  • test.support.changecwd(_path, quiet=False)
  • A context manager that temporarily changes the current workingdirectory to path and yields the directory.

If quiet is False, the context manager raises an exceptionon error. Otherwise, it issues only a warning and keeps the currentworking directory the same.

  • test.support.tempcwd(_name='tempcwd', quiet=False)
  • A context manager that temporarily creates a new directory andchanges the current working directory (CWD).

The context manager creates a temporary directory in the currentdirectory with name name before temporarily changing the currentworking directory. If name is None, the temporary directory iscreated using tempfile.mkdtemp().

If quiet is False and it is not possible to create or changethe CWD, an error is raised. Otherwise, only a warning is raisedand the original CWD is used.

  • test.support.tempumask(_umask)
  • A context manager that temporarily sets the process umask.
  • test.support.transientinternet(_resource_name, *, timeout=30.0, errnos=())
  • A context manager that raises ResourceDenied when various issueswith the internet connection manifest themselves as exceptions.
  • test.support.disable_faulthandler()
  • A context manager that replaces sys.stderr with sys.stderr.
  • test.support.gc_collect()
  • Force as many objects as possible to be collected. This is needed becausetimely deallocation is not guaranteed by the garbage collector. This meansthat del methods may be called later than expected and weakrefsmay remain alive for longer than expected.
  • test.support.disable_gc()
  • A context manager that disables the garbage collector upon entry andreenables it upon exit.
  • test.support.swapattr(_obj, attr, new_val)
  • Context manager to swap out an attribute with a new object.

用法:

  1. with swap_attr(obj, "attr", 5):
  2. ...

This will set obj.attr to 5 for the duration of the with block,restoring the old value at the end of the block. If attr doesn'texist on obj, it will be created and then deleted at the end of theblock.

The old value (or None if it doesn't exist) will be assigned to thetarget of the "as" clause, if there is one.

  • test.support.swapitem(_obj, attr, new_val)
  • Context manager to swap out an item with a new object.

用法:

  1. with swap_item(obj, "item", 5):
  2. ...

This will set obj["item"] to 5 for the duration of the with block,restoring the old value at the end of the block. If item doesn'texist on obj, it will be created and then deleted at the end of theblock.

The old value (or None if it doesn't exist) will be assigned to thetarget of the "as" clause, if there is one.

  • test.support.waitthreads_exit(_timeout=60.0)
  • Context manager to wait until all threads created in the with statementexit.
  • test.support.startthreads(_threads, unlock=None)
  • Context manager to start threads. It attempts to join the threads uponexit.
  • test.support.calcobjsize(fmt)
  • Return struct.calcsize() for nP{fmt}0n or, if gettotalrefcountexists, 2PnP{fmt}0P.
  • test.support.calcvobjsize(fmt)
  • Return struct.calcsize() for nPn{fmt}0n or, if gettotalrefcountexists, 2PnPn{fmt}0P.
  • test.support.checksizeof(test, o, size)
  • For testcase test, assert that the sys.getsizeof for o plus the GCheader size equals size.
  • test.support.can_symlink()
  • Return True if the OS supports symbolic links, Falseotherwise.
  • test.support.can_xattr()
  • Return True if the OS supports xattr, Falseotherwise.
  • @test.support.skip_unless_symlink
  • A decorator for running tests that require support for symbolic links.
  • @test.support.skip_unless_xattr
  • A decorator for running tests that require support for xattr.
  • @test.support.skip_unless_bind_unix_socket
  • A decorator for running tests that require a functional bind() for Unixsockets.
  • @test.support.anticipatefailure(_condition)
  • A decorator to conditionally mark tests withunittest.expectedFailure(). Any use of this decorator shouldhave an associated comment identifying the relevant tracker issue.
  • @test.support.runwith_locale(_catstr, *locales)
  • A decorator for running a function in a different locale, correctlyresetting it after it has finished. catstr is the locale category asa string (for example "LCALL"). The _locales passed will be triedsequentially, and the first valid locale will be used.
  • @test.support.runwith_tz(_tz)
  • A decorator for running a function in a specific timezone, correctlyresetting it after it has finished.
  • @test.support.requiresfreebsd_version(*minversion)
  • Decorator for the minimum version when running test on FreeBSD. If theFreeBSD version is less than the minimum, raise unittest.SkipTest.
  • @test.support.requireslinux_version(*minversion)
  • Decorator for the minimum version when running test on Linux. If theLinux version is less than the minimum, raise unittest.SkipTest.
  • @test.support.requiresmac_version(*minversion)
  • Decorator for the minimum version when running test on Mac OS X. If theMAC OS X version is less than the minimum, raise unittest.SkipTest.
  • @test.support.requires_IEEE_754
  • Decorator for skipping tests on non-IEEE 754 platforms.
  • @test.support.requires_zlib
  • Decorator for skipping tests if zlib doesn't exist.
  • @test.support.requires_gzip
  • Decorator for skipping tests if gzip doesn't exist.
  • @test.support.requires_bz2
  • Decorator for skipping tests if bz2 doesn't exist.
  • @test.support.requires_lzma
  • Decorator for skipping tests if lzma doesn't exist.
  • @test.support.requiresresource(_resource)
  • Decorator for skipping tests if resource is not available.
  • @test.support.requires_docstrings
  • Decorator for only running the test if HAVE_DOCSTRINGS.
  • @test.support.cpythononly(_test)
  • Decorator for tests only applicable to CPython.
  • @test.support.impldetail(_msg=None, **guards)
  • Decorator for invoking check_impl_detail() on guards. If thatreturns False, then uses msg as the reason for skipping the test.
  • @test.support.notracing(_func)
  • Decorator to temporarily turn off tracing for the duration of the test.
  • @test.support.refcounttest(_test)
  • Decorator for tests which involve reference counting. The decorator doesnot run the test if it is not run by CPython. Any trace function is unsetfor the duration of the test to prevent unexpected refcounts caused bythe trace function.
  • @test.support.reapthreads(_func)
  • Decorator to ensure the threads are cleaned up even if the test fails.
  • @test.support.bigmemtest(size, memuse, dry_run=True)
  • Decorator for bigmem tests.

size is a requested size for the test (in arbitrary, test-interpretedunits.) memuse is the number of bytes per unit for the test, or a goodestimate of it. For example, a test that needs two byte buffers, of 4 GiBeach, could be decorated with @bigmemtest(size=_4G, memuse=2).

The size argument is normally passed to the decorated test method as anextra argument. If dry_run is True, the value passed to the testmethod may be less than the requested value. If dry_run is False, itmeans the test doesn't support dummy runs when -M is not specified.

  • @test.support.bigaddrspacetest(f)
  • Decorator for tests that fill the address space. f is the function towrap.
  • test.support.make_bad_fd()
  • Create an invalid file descriptor by opening and closing a temporary file,and returning its descriptor.
  • test.support.checksyntax_error(_testcase, statement, errtext='', *, lineno=None, offset=None)
  • Test for syntax errors in statement by attempting to compile statement.testcase is the unittest instance for the test. errtext is thetext of the error raised by SyntaxError. If lineno is not None,compares to the line of the SyntaxError. If offset is not None,compares to the offset of the SyntaxError.
  • test.support.openurlresource(_url, *args, **kw)
  • Open url. If open fails, raises TestFailed.
  • test.support.importmodule(_name, deprecated=False, *, required_on())
  • This function imports and returns the named module. Unlike a normalimport, this function raises unittest.SkipTest if the modulecannot be imported.

Module and package deprecation messages are suppressed during this importif deprecated is True. If a module is required on a platform butoptional for others, set required_on to an iterable of platform prefixeswhich will be compared against sys.platform.

3.1 新版功能.

  • test.support.importfresh_module(_name, fresh=(), blocked=(), deprecated=False)
  • This function imports and returns a fresh copy of the named Python moduleby removing the named module from sys.modules before doing the import.Note that unlike reload(), the original module is not affected bythis operation.

fresh is an iterable of additional module names that are also removedfrom the sys.modules cache before doing the import.

blocked is an iterable of module names that are replaced with Nonein the module cache during the import to ensure that attempts to importthem raise ImportError.

The named module and any modules named in the fresh and _blocked_parameters are saved before starting the import and then reinserted intosys.modules when the fresh import is complete.

Module and package deprecation messages are suppressed during this importif deprecated is True.

This function will raise ImportError if the named module cannot beimported.

Example use:

  1. # Get copies of the warnings module for testing without affecting the
  2. # version being used by the rest of the test suite. One copy uses the
  3. # C implementation, the other is forced to use the pure Python fallback
  4. # implementation
  5. py_warnings = import_fresh_module('warnings', blocked=['_warnings'])
  6. c_warnings = import_fresh_module('warnings', fresh=['_warnings'])

3.1 新版功能.

  • test.support.modules_setup()
  • Return a copy of sys.modules.
  • test.support.modulescleanup(_oldmodules)
  • Remove modules except for oldmodules and encodings in order topreserve internal cache.
  • test.support.threading_setup()
  • Return current thread count and copy of dangling threads.
  • test.support.threadingcleanup(*originalvalues)
  • Cleanup up threads not specified in original_values. Designed to emita warning if a test leaves running threads in the background.
  • test.support.jointhread(_thread, timeout=30.0)
  • Join a thread within timeout. Raise an AssertionError if threadis still alive after timeout seconds.
  • test.support.reap_children()
  • Use this at the end of test_main whenever sub-processes are started.This will help ensure that no extra children (zombies) stick around tohog resources and create problems when looking for refleaks.
  • test.support.bindport(_sock, host=HOST)
  • Bind the socket to a free port and return the port number. Relies onephemeral ports in order to ensure we are using an unbound port. This isimportant as many tests may be running simultaneously, especially in abuildbot environment. This method raises an exception if thesock.family is AF_INET and sock.type isSOCK_STREAM, and the socket hasSO_REUSEADDR or SO_REUSEPORT set on it.Tests should never set these socket options for TCP/IP sockets.The only case for setting these options is testing multicasting viamultiple UDP sockets.

Additionally, if the SO_EXCLUSIVEADDRUSE socket option isavailable (i.e. on Windows), it will be set on the socket. This willprevent anyone else from binding to our host/port for the duration of thetest.

  • test.support.findunused_port(_family=socket.AF_INET, socktype=socket.SOCK_STREAM)
  • Returns an unused port that should be suitable for binding. This isachieved by creating a temporary socket with the same family and type asthe sock parameter (default is AF_INET,SOCK_STREAM),and binding it to the specified host address (defaults to 0.0.0.0)with the port set to 0, eliciting an unused ephemeral port from the OS.The temporary socket is then closed and deleted, and the ephemeral port isreturned.

Either this method or bind_port() should be used for any testswhere a server socket needs to be bound to a particular port for theduration of the test.Which one to use depends on whether the calling code is creating a Pythonsocket, or if an unused port needs to be provided in a constructoror passed to an external program (i.e. the -accept argument toopenssl's s_server mode). Always prefer bind_port() overfind_unused_port() where possible. Using a hard coded port isdiscouraged since it can make multiple instances of the test impossible torun simultaneously, which is a problem for buildbots.

  • test.support.loadpackage_tests(_pkg_dir, loader, standard_tests, pattern)
  • Generic implementation of the unittest loadtests protocol foruse in test packages. _pkg_dir is the root directory of the package;loader, standard_tests, and pattern are the arguments expected byloadtests. In simple cases, the test package's _init.pycan be the following:
  1. import os
  2. from test.support import load_package_tests
  3.  
  4. def load_tests(*args):
  5. return load_package_tests(os.path.dirname(__file__), *args)
  • test.support.fsis_case_insensitive(_directory)
  • Return True if the file system for directory is case-insensitive.
  • test.support.detectapi_mismatch(_ref_api, other_api, *, ignore=())
  • Returns the set of attributes, functions or methods of ref_api notfound on other_api, except for a defined list of items to beignored in this check specified in ignore.

By default this skips private attributes beginning with '' butincludes all magic methods, i.e. those starting and ending in '_'.

3.5 新版功能.

  • test.support.patch(test_instance, object_to_patch, attr_name, new_value)
  • Override object_to_patch.attr_name with new_value. Also addcleanup procedure to test_instance to restore object_to_patch forattr_name. The attr_name should be a valid attribute forobject_to_patch.
  • test.support.checkfree_after_iterating(_test, iter, cls, args=())
  • Assert that iter is deallocated after iterating.
  • test.support.missingcompiler_executable(_cmd_names=[])
  • Check for the existence of the compiler executables whose names are listedin cmd_names or all the compiler executables when cmd_names is emptyand return the first missing executable or None when none is foundmissing.
  • test.support.checkall(test_case, module, name_of_module=None, extra=(), blacklist=())
  • Assert that the all variable of module contains all public names.

The module's public names (its API) are detected automaticallybased on whether they match the public name convention and were defined inmodule.

The name_of_module argument can specify (as a string or tuple thereof) whatmodule(s) an API could be defined in order to be detected as a publicAPI. One case for this is when module imports part of its public API fromother modules, possibly a C backend (like csv and its _csv).

The extra argument can be a set of names that wouldn't otherwise be automaticallydetected as "public", like objects without a proper moduleattribute. If provided, it will be added to the automatically detected ones.

The blacklist argument can be a set of names that must not be treated as part ofthe public API even though their names indicate otherwise.

Example use:

  1. import bar
  2. import foo
  3. import unittest
  4. from test import support
  5.  
  6. class MiscTestCase(unittest.TestCase):
  7. def test__all__(self):
  8. support.check__all__(self, foo)
  9.  
  10. class OtherTestCase(unittest.TestCase):
  11. def test__all__(self):
  12. extra = {'BAR_CONST', 'FOO_CONST'}
  13. blacklist = {'baz'} # Undocumented name.
  14. # bar imports part of its API from _bar.
  15. support.check__all__(self, bar, ('bar', '_bar'),
  16. extra=extra, blacklist=blacklist)

3.6 新版功能.

The test.support module defines the following classes:

  • class test.support.TransientResource(exc, **kwargs)
  • Instances are a context manager that raises ResourceDenied if thespecified exception type is raised. Any keyword arguments are treated asattribute/value pairs to be compared against any exception raised within thewith statement. Only if all pairs match properly againstattributes on the exception is ResourceDenied raised.
  • class test.support.EnvironmentVarGuard
  • Class used to temporarily set or unset environment variables. Instances canbe used as a context manager and have a complete dictionary interface forquerying/modifying the underlying os.environ. After exit from thecontext manager all changes to environment variables done through thisinstance will be rolled back.

在 3.1 版更改: Added dictionary interface.

  • EnvironmentVarGuard.set(envvar, value)
  • Temporarily set the environment variable envvar to the value ofvalue.
  • EnvironmentVarGuard.unset(envvar)
  • Temporarily unset the environment variable envvar.
  • class test.support.SuppressCrashReport
  • A context manager used to try to prevent crash dialog popups on tests thatare expected to crash a subprocess.

On Windows, it disables Windows Error Reporting dialogs usingSetErrorMode.

On UNIX, resource.setrlimit() is used to setresource.RLIMIT_CORE's soft limit to 0 to prevent coredump filecreation.

On both platforms, the old value is restored by exit().

  • class test.support.CleanImport(*module_names)
  • A context manager to force import to return a new module reference. Thisis useful for testing module-level behaviors, such as the emission of aDeprecationWarning on import. Example usage:
  1. with CleanImport('foo'):
  2. importlib.import_module('foo') # New reference.
  • class test.support.DirsOnSysPath(*paths)
  • A context manager to temporarily add directories to sys.path.

This makes a copy of sys.path, appends any directories givenas positional arguments, then reverts sys.path to the copiedsettings when the context ends.

Note that allsys.path modifications in the body of thecontext manager, including replacement of the object,will be reverted at the end of the block.

  • class test.support.SaveSignals
  • Class to save and restore signal handlers registered by the Python signalhandler.
  • class test.support.Matcher
    • matches(self, d, **kwargs)
    • Try to match a single dict with the supplied arguments.

    • matchvalue(_self, k, dv, v)

    • Try to match a single stored value (dv) with a supplied value (v).
  • class test.support.WarningsRecorder
  • Class used to record warnings for unit tests. See documentation ofcheck_warnings() above for more details.
  • class test.support.BasicTestRunner
    • run(test)
    • Run test and return the result.
  • class test.support.TestHandler(logging.handlers.BufferingHandler)
  • Class for logging support.
  • class test.support.FakePath(path)
  • Simple path-like object. It implements the fspath()method which just returns the path argument. If path is an exception,it will be raised in fspath().

test.support.script_helper —- Utilities for the Python execution tests

The test.support.script_helper module provides support for Python'sscript execution tests.

  • test.support.script_helper.interpreter_requires_environment()
  • Return True if sys.executable interpreter requires environmentvariables in order to be able to run at all.

This is designed to be used with @unittest.skipIf() to annotate teststhat need to use an assert_python*() function to launch an isolatedmode (-I) or no environment mode (-E) sub-interpreter process.

A normal build & test does not run into this situation but it can happenwhen trying to run the standard library test suite from an interpreter thatdoesn't have an obvious home with Python's current home finding logic.

Setting PYTHONHOME is one way to get most of the testsuite to runin that situation. PYTHONPATH or PYTHONUSERSITE areother common environment variables that might impact whether or not theinterpreter can start.

  • test.support.scripthelper.run_python_until_end(args, *envvars)
  • Set up the environment based on env_vars for running the interpreterin a subprocess. The values can include isolated, cleanenv,__cwd, and TERM.
  • test.support.scripthelper.assert_python_ok(args, *envvars)
  • Assert that running the interpreter with args and optional environmentvariables env_vars succeeds (rc == 0) and return a (return code,stdout, stderr) tuple.

If the _cleanenv keyword is set, _env_vars is used as a freshenvironment.

Python is started in isolated mode (command line option -I),except if the __isolated keyword is set to False.

  • test.support.scripthelper.assert_python_failure(args, *envvars)
  • Assert that running the interpreter with args and optional environmentvariables env_vars fails (rc != 0) and return a (return code,stdout, stderr) tuple.

See assert_python_ok() for more options.

  • test.support.scripthelper.spawn_python(args, _stdout=subprocess.PIPE, stderr=subprocess.STDOUT, _*kw_)
  • Run a Python subprocess with the given arguments.

kw is extra keyword args to pass to subprocess.Popen(). Returns asubprocess.Popen object.

  • test.support.scripthelper.kill_python(_p)
  • Run the given subprocess.Popen process until completion and returnstdout.
  • test.support.scripthelper.make_script(_script_dir, script_basename, source, omit_suffix=False)
  • Create script containing source in path script_dir and script_basename.If omit_suffix is False, append .py to the name. Return the fullscript path.
  • test.support.scripthelper.make_zip_script(_zip_dir, zip_basename, script_name, name_in_zip=None)
  • Create zip file at zip_dir and zip_basename with extension zip whichcontains the files in script_name. name_in_zip is the archive name.Return a tuple containing (full path, full path of archive name).
  • test.support.scripthelper.make_pkg(_pkg_dir, init_source='')
  • Create a directory named pkg_dir containing an init file withinit_source as its contents.
  • test.support.scripthelper.make_zip_pkg(_zip_dir, zip_basename, pkg_name, script_basename, source, depth=1, compiled=False)
  • Create a zip package directory with a path of zip_dir and zipbasenamecontaining an empty __init file and a file _script_basename_containing the _source. If compiled is True, both source files willbe compiled and added to the zip package. Return a tuple of the full zippath and the archive name for the zip file.