Google Test C++ Unit Testing

Setup

NOTE: Setup of unit tests assumes you have the appropriate setup for running googletest.

Python and pip must be on path to install and use virtual environments for report generation.

Run a git submodule update --init --recursive to ensure that the googletest submodule is initialized properly to be built with the application.

Add Test Cases

In order to add test cases, simply add a TEST() macro call. If adding a new suite / file. Add a folder for the interface ex: interface/ for interface_file.cpp and then your test file name should be your test name with test_*.cc ex: test_interface_file.cc. If you have a test fixture be sure to use TEST_F() instead.

Running Tests

From the project directory, run make test to run all tests available to google test. This will also automatically generate the coverage report as well. The tests will store the output in a log file in the build folder. To run the tests with the gtest output, run make gtest instead. This will produce the output from googletest directly with all printouts from the tests and source code.

In order to run specific tests, specify the CASES macro in your command.
Set the value to be your test cases and fixtures separated by : for example:

make test CASES=TestFixture.* - runs all of the tests inside of the fixture named TestFixture.

make test CASES=TestFixture.TestCaseName - runs one test named TestCaseName inside of the fixture named TestFixture.

make test CASES=TestFixture.TestCaseName:TestFixture.TestCaseName - runs two test cases, TestCaseName and TestCaseName, inside of the fixture named TestFixture.

make test CASES=TestFixture1.*:TestFixture2.* - runs all of the tests inside of two fixtures, TestFixture1 and TestFixture2.

Report Generation

Report generation with gcovr happens at the end of the make test target, but can be run individually with a make report command. Within the build directory, a gcov folder will contain the html output. The overall results will be in the file {project_name}_test.html and be nested to expand into the results for each source file's coverage report. Gcovr 7.0 is used in a python venv activated within the report generating target.

Mock Generation

A simple gmock generator is in the scripts folder and can be invoked with the make mock FILE=... target where FILE is set to the path to the desired C++ header to mock. Once generated, the mock class is output automatically to the ./test/mocks directory. Know that the script has limitations when it comes to the nesting of namespaces and classes in C++, and that mock classes may not be completely accurate and need to be reviewed.

Memory Sanitization

Unit tests are compiled with the -fsanitize=address flag which uses AddressSanitizer to check for memory mismanagement. This runs on linux builds so ensure that you have the sanatizer installed if needed. This will also fail unit tests if memory is not handled properly in tests.