This is a collection of quite useful scripts that expand the possibilities for building software with CMake, by making some things easier and otherwise adding new build types
- C++ Standards
c++-standards.cmake - Sanitizer Builds
sanitizers.cmake - Code Coverage
code-coverage.cmake - Compiler Options
compiler-options.cmake - Dependency Graph
dependency-graph.cmake - Doxygen
doxygen.cmake - Tools
tools.cmake - Formatting
formatting.cmake
C++ Standards c++-standards.cmake
Using the functions cxx_11(), cxx_14(), cxx_17() or cxx_20() this adds the appropriated flags for both unix and MSVC compilers, even for those before 3.11 with improper support.
These obviously force the standard to be required, and also disables compiler-specific extensions, ie --std=gnu++11. This helps to prevent fragmenting the code base with items not available elsewhere, adhering to the agreed C++ standards only.
Sanitizer Builds sanitizers.cmake
Sanitizers are tools that perform checks during a program’s runtime and returns issues, and as such, along with unit testing, code coverage and static analysis, is another tool to add to the programmers toolbox. And of course, like the previous tools, are tragically simple to add into any project using CMake, allowing any project and developer to quickly and easily use.
A quick rundown of the tools available, and what they do:
- LeakSanitizer detects memory leaks, or issues where memory is allocated and never deallocated, causing programs to slowly consume more and more memory, eventually leading to a crash.
- AddressSanitizer is a fast memory error detector. It is useful for detecting most issues dealing with memory, such as:
- Out of bounds accesses to heap, stack, global
- Use after free
- Use after return
- Use after scope
- Double-free, invalid free
- Memory leaks (using LeakSanitizer)
- ThreadSanitizer detects data races for multi-threaded code.
- UndefinedBehaviourSanitizer detects the use of various features of C/C++ that are explicitly listed as resulting in undefined behaviour. Most notably:
- Using misaligned or null pointer.
- Signed integer overflow
- Conversion to, from, or between floating-point types which would overflow the destination
- Division by zero
- Unreachable code
- MemorySanitizer detects uninitialized reads.
These are used by declaring the USE_SANITIZER CMake variable as one of:
- Address
- Memory
- MemoryWithOrigins
- Undefined
- Thread
- Address;Undefined
- Undefined;Address
- Leak
Code Coverage code-coverage.cmake
In computer science, test coverage is a measure used to describe the degree to which the source code of a program is executed when a particular test suite runs. A program with high test coverage, measured as a percentage, has had more of its source code executed during testing, which suggests it has a lower chance of containing undetected software bugs compared to a program with low test coverage. Many different metrics can be used to calculate test coverage; some of the most basic are the percentage of program subroutines and the percentage of program statements called during execution of the test suite.
Code coverage is the detailing of, during the execution of a binary, which regions, functions, or lines of code are actually executed. This can be used in a number of ways, from figuring out areas that automated testing is lacking or not touching, to giving a user an instrumented binary to determine which areas of code are used most/least to determine which areas to focus on. Although this does come with the caveat that coverage is no guarantee of good testing, just of what code has been.
Coverage here is supported on both GCC and Clang. GCC requires the lcov program, and Clang requires llvm-cov and llvm-profdata, often provided with the llvm toolchain.
To enable, turn on the CODE_COVERAGE variable.
- GCOV/LCOV:
- ccov : Generates HTML code coverage report for every target added with 'AUTO' parameter.
- ccov-${TARNGET_NAME} : Generates HTML code coverage report for the associated named target.
- ccov-all : Generates HTML code coverage report, merging every target added with 'ALL' parameter into a single detailed report.
- LLVM-COV:
- ccov : Generates HTML code coverage report for every target added with 'AUTO' parameter.
- ccov-report : Generates HTML code coverage report for every target added with 'AUTO' parameter.
- ccov-${TARGET_NAME} : Generates HTML code coverage report.
- ccov-rpt-${TARGET_NAME} : Prints to command line summary per-file coverage information.
- ccov-show-${TARGET_NAME} : Prints to command line detailed per-line coverage information.
- ccov-all : Generates HTML code coverage report, merging every target added with 'ALL' parameter into a single detailed report.
- ccov-all-report : Prints summary per-file coverage information for every target added with ALL' parameter to the command line.
To enable any code coverage instrumentation/targets, the single CMake option of CODE_COVERAGE needs to be set to 'ON', either by GUI, ccmake, or on the command line ie -DCODE_COVERAGE=ON.
From this point, there are two primary methods for adding instrumentation to targets:
- A blanket instrumentation by calling
add_code_coverage(), where all targets in that directory and all subdirectories are automatically instrumented. - Per-target instrumentation by calling
target_code_coverage(<TARGET_NAME>), where the target is given and thus only that target is instrumented. This applies to both libraries and executables.
To add coverage targets, such as calling make ccov to generate the actual coverage information for perusal or consumption, call target_code_coverage(<TARGET_NAME>) on an executable target.
NOTE: For more options, please check the actual code-coverage.cmake file.
In this case, the coverage information reported will will be that of the theLib library target and theExe executable.
add_code_coverage() # Adds instrumentation to all targets
add_library(theLib lib.cpp)
add_executable(theExe main.cpp)
target_link_libraries(theExe PRIVATE theLib)
target_code_coverage(theExe) # As an executable target, adds the 'ccov-theExe' target (instrumentation already added via global anyways) for generating code coverage reports.
add_library(theLib lib.cpp)
target_code_coverage(theLib) # As a library target, adds coverage instrumentation but no targets.
add_executable(theExe main.cpp)
target_link_libraries(theExe PRIVATE theLib)
target_code_coverage(theExe) # As an executable target, adds the 'ccov-theExe' target and instrumentation for generating code coverage reports.
add_executable(theExe main.cpp non_covered.cpp)
target_code_coverage(theExe EXCLUDE non_covered.cpp) # As an executable target, the reports will exclude the non-covered.cpp file.
add_code_coverage_all_targets(EXCLUDE test/*) # Adds the 'ccov-all' target set and sets it to exclude all files in test/ folders.
add_executable(theExe main.cpp non_covered.cpp)
target_code_coverage(theExe AUTO ALL EXCLUDE non_covered.cpp test/*) # As an executable target, adds to the 'ccov' and ccov-all' targets, and the reports will exclude the non-covered.cpp file, and any files in a test/ folder.
Compiler Options compiler-options.cmake
Allows for easy use of some pre-made compiler options for the major compilers.
Using -DENABLE_ALL_WARNINGS=ON will enable almost all of the warnings available for a compiler:
| Compiler | Options |
|---|---|
| MSVC | /W4 |
| GCC | -Wall -Wextra |
| Clang | -Wall -Wextra |
Using -DENABLE_EFFECTIVE_CXX=ON adds the -Weffc++ for both GCC and clang.
Dependency Graph dependency-graph.cmake
CMake, with the dot application available, will build a visual representation of the library/executable dependencies, like so:
OUTPUT_TYPE STR - The type of output of dot to produce. Can be whatever dot itself supports (eg. png, ps, pdf).
ADD_TO_DEP_GRAPH - If specified, add this generated target to be a dependency of the more general dep-graph target.
TARGET_NAME STR - The name to give the doc target. (Default: doc-${PROJECT_NAME})
OUTPUT_DIR STR - The directory to place the generated output
Doxygen doxygen.cmake
Builds doxygen documentation with a default 'Doxyfile.in' or with a specified one, and can make the results installable (under the doc install target)
This can only be used once per project, as each target generated is as doc-${PROJECT_NAME} unless TARGET_NAME is specified.
ADD_TO_DOC - If specified, adds this generated target to be a dependency of the more general doc target.
INSTALLABLE - Adds the generated documentation to the generic install target, under the documentation installation group.
PROCESS_DOXYFILE - If set, then will process the found Doxyfile through the CMAKE configure_file function for macro replacements before using it. (@ONLY)
TARGET_NAME STR - The name to give the doc target. (Default: doc-${PROJECT_NAME})
OUTPUT_DIR STR - The directory to place the generated output. (Default: ${CMAKE_CURRENT_BINARY_DIR}/doc)
INSTALL_PATH STR - The path to install the documenttation under. (if not specified, defaults to 'share/${PROJECT_NAME})
DOXYFILE_PATH STR - The given doxygen file to use/process. (Defaults to'${CMAKE_CURRENT_SOURCE_DIR}/Doxyfile')
Tools tools.cmake
clang-tidy is a clang-based C++ “linter” tool. Its purpose is to provide an extensible framework for diagnosing and fixing typical programming errors, like style violations, interface misuse, or bugs that can be deduced via static analysis. clang-tidy is modular and provides a convenient interface for writing new checks.
When detected, clang-tidy can be enabled by using the option of -DCLANG_TIDY=ON, as it is disabled by default.
To use, add the clang_tidy() function, with the arguments being the options to pass to the clang tidy program, such as '-checks=*'.
This tool helps to organize headers for all files encompass all items being used in that file, without accidentally relying upon headers deep down a chain of other headers. This is disabled by default, and can be enabled via have the program installed and adding -DIWYU=ON.
To use, add the include_what_you_use() function, with the arguments being the options to pass to the program.
This tool is another static analyzer in the vein of clang-tidy, which focuses on having no false positives. This is by default disabled, and can be enabled via have the program installed and adding -DCPPCHECK=ON.
To use, add the cppcheck() function, with the arguments being the options to pass to the program.
Formatting formatting.cmake
Allows to automatically perform code formatting using the clang-format program, by calling an easy-to-use target ala make format. It requires a target name, and the list of files to format. As well, if the target name is the name of another target, then all files associated with that target will be added, and the target name changed to be format_<TARGET>. As well, any targets otherwise listed with the files will also have their files imported for formatting.
file(GLOB_RECURSE ALL_CODE_FILES
${PROJECT_SOURCE_DIR}/src/*.[ch]pp
${PROJECT_SOURCE_DIR}/src/*.[ch]
${PROJECT_SOURCE_DIR}/include/*.[h]pp
${PROJECT_SOURCE_DIR}/include/*.[h]
${PROJECT_SOURCE_DIR}/example/*.[ch]pp
${PROJECT_SOURCE_DIR}/example/*.[ch]
)
clang_format(TARGET_NAME ${ALL_CODE_FILES})
Similar to the clang-format above, creates a target cmake-format when the cmake_format(<FILES>) function is defined in CMake scripts, and any passed in will be formatted by the cmake-format program, if it is found.
file(GLOB_RECURSE CMAKE_FILES
CMakeLists.txt
)
cmake_format(TARGET_NAME ${CMAKE_FILES})