Skip to content

Latest commit

 

History

History
162 lines (129 loc) · 4.99 KB

4.9-chinese.md

File metadata and controls

162 lines (129 loc) · 4.99 KB

4.9 运行测试子集

NOTE:此示例代码可以在 https://github.com/dev-cafe/cmake-cookbook/tree/v1.0/chapter-04/recipe-09 中找到。该示例在CMake 3.5版(或更高版本)中是有效的,并且已经在GNU/Linux、macOS和Windows上进行过测试。

前面的示例中,我们学习了如何在CMake的帮助下并行运行测试,并讨论了从最长的测试开始是最高效的。虽然,这种策略将总测试时间最小化,但是在特定特性的代码开发期间,或者在调试期间,我们可能不希望运行整个测试集。对于调试和代码开发,我们只需要能够运行选定的测试子集。在本示例中,我们将实现这一策略。

准备工作

在这个例子中,我们假设总共有六个测试:前三个测试比较短,名称分别为feature-afeature-bfeature-c,还有三个长测试,名称分别是feature-dbenchmark-abenchmark-b。这个示例中,我们可以用Python脚本表示这些测试,可以在其中调整休眠时间:

import sys
import time

# wait for 0.1 seconds
time.sleep(0.1)

# finally report success
sys.exit(0)

具体实施

以下是我们CMakeLists.txt文件内容的详细内容:

  1. CMakeLists.txt中,定义了六个测试:

    cmake_minimum_required(VERSION 3.5 FATAL_ERROR)
    
    # project name
    project(recipe-09 LANGUAGES NONE)
    
    # detect python
    find_package(PythonInterp REQUIRED)
    
    # define tests
    enable_testing()
    
    add_test(
      NAME feature-a
      COMMAND ${PYTHON_EXECUTABLE} ${CMAKE_CURRENT_SOURCE_DIR}/test/feature-a.py
      )
    add_test(
      NAME feature-b
      COMMAND ${PYTHON_EXECUTABLE} ${CMAKE_CURRENT_SOURCE_DIR}/test/feature-b.py
      )
    add_test(
      NAME feature-c
      COMMAND ${PYTHON_EXECUTABLE} ${CMAKE_CURRENT_SOURCE_DIR}/test/feature-c.py
      )
    add_test(
      NAME feature-d
      COMMAND ${PYTHON_EXECUTABLE} ${CMAKE_CURRENT_SOURCE_DIR}/test/feature-d.py
      )
    add_test(
      NAME benchmark-a
      COMMAND ${PYTHON_EXECUTABLE} ${CMAKE_CURRENT_SOURCE_DIR}/test/benchmark-a.py
      )
    add_test(
      NAME benchmark-b
      COMMAND ${PYTHON_EXECUTABLE} ${CMAKE_CURRENT_SOURCE_DIR}/test/benchmark-b.py
      )
  2. 此外,我们给较短的测试贴上quick的标签,给较长的测试贴上long的标签:

    set_tests_properties(
      feature-a
      feature-b
      feature-c
      PROPERTIES
      	LABELS "quick"
      )
    set_tests_properties(
      feature-d
      benchmark-a
      benchmark-b
      PROPERTIES
      	LABELS "long"
      )
  3. 我们现在可以运行测试集了,如下:

    $ mkdir -p build
    $ cd build
    $ cmake ..
    $ ctest
    
    Start 1: feature-a
    1/6 Test #1: feature-a ........................ Passed 0.11 sec
    Start 2: feature-b
    2/6 Test #2: feature-b ........................ Passed 0.11 sec
    Start 3: feature-c
    3/6 Test #3: feature-c ........................ Passed 0.11 sec
    Start 4: feature-d
    4/6 Test #4: feature-d ........................ Passed 0.51 sec
    Start 5: benchmark-a
    5/6 Test #5: benchmark-a ...................... Passed 0.51 sec
    Start 6: benchmark-b
    6/6 Test #6: benchmark-b ...................... Passed 0.51 sec
    100% tests passed, 0 tests failed out of 6
    Label Time Summary:
    long = 1.54 sec*proc (3 tests)
    quick = 0.33 sec*proc (3 tests)
    Total Test time (real) = 1.87 sec

工作原理

现在每个测试都有一个名称和一个标签。CMake中所有的测试都是有编号的,所以它们也带有唯一编号。定义了测试标签之后,我们现在可以运行整个集合,或者根据它们的名称(使用正则表达式)、标签或编号运行测试。

按名称运行测试(运行所有具有名称匹配功能的测试):

$ ctest -R feature

Start 1: feature-a
1/4 Test #1: feature-a ........................ Passed 0.11 sec
Start 2: feature-b
2/4 Test #2: feature-b ........................ Passed 0.11 sec
Start 3: feature-c
3/4 Test #3: feature-c ........................ Passed 0.11 sec
Start 4: feature-d
4/4 Test #4: feature-d ........................ Passed 0.51 sec
100% tests passed, 0 tests failed out of 4

按照标签运行测试(运行所有的长测试):

$ ctest -L long

Start 4: feature-d
1/3 Test #4: feature-d ........................ Passed 0.51 sec
Start 5: benchmark-a
2/3 Test #5: benchmark-a ...................... Passed 0.51 sec
Start 6: benchmark-b
3/3 Test #6: benchmark-b ...................... Passed 0.51 sec
100% tests passed, 0 tests failed out of 3

根据数量运行测试(运行测试2到4)产生的结果是:

$ ctest -I 2,4

Start 2: feature-b
1/3 Test #2: feature-b ........................ Passed 0.11 sec
Start 3: feature-c
2/3 Test #3: feature-c ........................ Passed 0.11 sec
Start 4: feature-d
3/3 Test #4: feature-d ........................ Passed 0.51 sec
100% tests passed, 0 tests failed out of 3

更多信息

尝试使用$ ctest --help,将看到有大量的选项可供用来定制测试。