使用测试异步库的功能增强标准unittest包
asynctest的Python项目详细描述
异步测试
包asynctest构建在标准unittest模块之上,并且 在测试异步库时减少样板代码。
目前,AsyncTest的目标是“选择器”模型,因此,有些特性 不会吧?使用windows的代理。
警告
由于AsyncTest0.13,因此不再支持Python3.4。
功能
测试用例
- Initialize and close a loop created for each test (it can be
configurated), if the loop uses a selector, it will be updated with
a TestSelector object wrapping the original selector (see below),
- if the test function is a coroutine function or returns a coroutine, it
will run on the loop,
- TestCase.setUp() and TestCase.tearDown() can be coroutine functions,
- control post-test checks with @fail_on, for instance, the test fail if
the loop didn’t run, some optional checks can be activated,
- ClockedTestCase allows to control the loop clock and run timed events
without waiting the wall clock.
- Initialize and close a loop created for each test (it can be configurated), if the loop uses a selector, it will be updated with a TestSelector object wrapping the original selector (see below),
- if the test function is a coroutine function or returns a coroutine, it will run on the loop,
- TestCase.setUp() and TestCase.tearDown() can be coroutine functions,
- control post-test checks with @fail_on, for instance, the test fail if the loop didn’t run, some optional checks can be activated,
- ClockedTestCase allows to control the loop clock and run timed events without waiting the wall clock.
模拟和协同工作
- CoroutineMock is a new Mock class which mocks a coroutine function, and
returns a coroutine when called,
- MagicMock supports asynchronous context managers and asynchronous
iterators,
- NonCallableMock, Mock and CoroutineMock can return CoroutineMock objects
when its attributes are get if there is a matching attribute in the spec
(or spec_set) object which is a coroutine function,
- patch(), patch.object(), patch.multiple() return a MagickMock or
CoroutineMock object by default, according to the patched target,
- patch(), patch.object(), patch.multiple() handle generators and coroutines
and their behavior can be controled when the generator or coroutine pauses,
- all the patch() methods can decorate coroutine functions,
- mock_open() returns a MagickMock object by default.
- return_once() can be used with Mock.side_effect to return a value only
once when a mock is called.
- CoroutineMock is a new Mock class which mocks a coroutine function, and returns a coroutine when called,
- MagicMock supports asynchronous context managers and asynchronous iterators,
- NonCallableMock, Mock and CoroutineMock can return CoroutineMock objects when its attributes are get if there is a matching attribute in the spec (or spec_set) object which is a coroutine function,
- patch(), patch.object(), patch.multiple() return a MagickMock or CoroutineMock object by default, according to the patched target,
- patch(), patch.object(), patch.multiple() handle generators and coroutines and their behavior can be controled when the generator or coroutine pauses,
- all the patch() methods can decorate coroutine functions,
- mock_open() returns a MagickMock object by default.
- return_once() can be used with Mock.side_effect to return a value only once when a mock is called.
选择器
模块asynctest.selector提供类来模拟执行IO的对象 (文件、套接字等)。
- FileMock is a special type of mock which represents a file. FileMock.fileno() returns a special value which allows to identify uniquely the mock,
- SocketMock is a special type of FileMock which uses socket.socket as spec,
- TestSelector is a custom selector able to wrap a real selector implementation and deal with FileMock objects, it can replace a selector loop by calling loop._selector = TestSelector(loop._selector), and will intercept mock so they don’t get registered to the actual selector.
- set_read_ready() and set_write_ready() to force read and write event callbacks to be scheduled on the loop, as if the selector scheduled them.
助手
- the coroutine exhaust_callbacks(loop) returns once all the callbacks which should be called immediately are executed, which is useful when the test author needs to assert things which are not yet executed by the loop.
路线图
我希望有时间开发和发布以下功能:
- 对常见错误的一组警告
- 演员支持
测试
AsyncTest是单元测试。您可以使用以下命令运行AsyncTest测试套件:
$ PYTHONPATH=. python -m unittest test