jenkinz / cmake-project-template

A C/C++ project template using modern CMake best practices.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

CMake Project Template (Linux C/C++ Application)

[[TOC]]

Overview

This is a template project for a Linux C/C++ library and/or application. It includes support for the following:

  • CMake build configurations
  • Formatting of source code for standards compliance with clang-format
  • Built-in static analysis with clang-tidy
  • Style checking with cpplint in accordance with the Google C++ Style Guide
  • Analysis of #include statements by include-what-you-use
  • Cyclomatic complexity checks per-function by pmccabe
  • Testing using CMake's ctest and the googletest framework
  • Documentation generation with doxygen
  • VS Code IDE support

The CMake project configuration works best utilizing the native host environment toolchain (i.e., not cross-compiling). Instead of cross-compiling, it's recommended to use the remote-ssh capabilities of VS Code (or your IDE of choice) to build and debug on the target directly.

This description should be modified when creating a new project based on this template.

Standards

Checks are included in this project template to adhere to the following code standards and guidelines, which the developer should review and be familiar with:

By extension from Autosar C++14, the following standards should also be reviewed:

  • MISRA C++:2008
  • Joint Strike Fighter Air Vehicle C++ Coding Standards
  • High Integrity C++ Coding Standard Version 4.0

Project Structure and Naming Conventions

The following naming conventions are used throughout this project:

  • Library source code implementation files are placed in src/
  • Library header interface files are placed in include/<library-name>/
    • In #include directives, always specify library name (e.g. #include "library-name/header.h")
  • Application source code and headers are placed in app-src/ (delete this folder when project is library-only)
  • Source code file names are all lowercase, separated by underscores (e.g. my_file.h or my_impl.c)
    • C++ implementation files use extension .cpp
    • C++-only header files use extension .hpp
    • C implementation files use extension .c
    • C header files, and header files intended to be shared between C++ and C, use extension .h
  • Target names are all lowercase; whenever possible, keep names to one word (e.g. mylib or myapp), but if separation is needed, use dashes (-) (e.g. my-lib or my-app)
    • Folder subdirectories for library includes should match the target name (e.g. include/libname/ or include/my-lib/)

Tool Prerequisites

The following tools should be installed on the host machine prior to building:

Optionally, you may also install the ninja build tool which can be used in leiu of make for slightly speedier builds.

macOS Host

Assuming you have Homebrew installed:

$ brew install ccache cmake cpplint doxygen graphviz include-what-you-use llvm pmccabe python
$ python3 -m pip install lcov_cobertura pre-commit
$ pre-commit install

You must then symlink the extra clang tools from the llvm installation:

$ ln -s "$(brew --prefix llvm)/bin/clang-format" "$(brew --prefix)/bin/clang-format"
$ ln -s "$(brew --prefix llvm)/bin/clang-tidy" "$(brew --prefix)/bin/clang-tidy"
$ ln -s "$(brew --prefix llvm)/bin/clang-apply-replacements" "$(brew --prefix)/bin/clang-apply-replacements"

Debian/Ubuntu Linux Host (incl. Windows WSL Debian/Ubuntu)

Note: LLVM version 13 is installed for the clang-extra tools including clang-tidy and clang-format. The default packages in Debian are too old (at version 7 as of this writing) so a newer version is installed below.

$ sudo apt update
$ sudo apt install build-essential doxygen graphviz gdb git ccache iwyu lcov ninja-build pmccabe python3-pip
$ wget https://github.com/Kitware/CMake/releases/download/v3.19.7/cmake-3.19.7.tar.gz
$ tar -xzf cmake-3.19.7.tar.gz
$ ./bootstrap -- -DCMAKE_BUILD_TYPE:STRING=Release
$ make
$ make install
$ rm -rf ./cmake-3.19.7*
$ wget https://apt.llvm.org/llvm.sh
$ chmod +x llvm.sh
$ sudo ./llvm.sh 13
$ sudo apt install clang-tidy-13 clang-format-13
$ sudo ln -sf $(which clang-tidy-13) /usr/local/bin/clang-tidy
$ sudo ln -sf $(which clang-format-13) /usr/local/bin/clang-format
$ sudo python3 -m pip install cpplint lcov_cobertura pre-commit
$ pre-commit install

Alpine Linux Host (incl. Windows WSL Alpine)

Note: need to become root to run the installation commands. Also, IWYU is not immediately available on Alpine.

$ su -
# apk update
# apk add alpine-sdk gdb git ccache cmake clang-extra-tools ninja py3-pip
# pip3 install cpplint lcov_cobertura pre-commit
# exit
$ pre-commit install

Configure

The usual CMake configuration procedure can be used from the project top-level. To use the recommended approach for configuration (that is, out-of-source builds within an unversioned build directory), run the following from the project top-level:

$ mkdir build
$ cd build
$ cmake ..

ninja may be used instead of make by using the following cmake command when configuring:

$ cmake -GNinja ..

Configuration Options

The following configuration options can be added to the cmake command (prepend each option with -D):

  • CMAKE_BUILD_TYPE=Debug|Release|RelWithDebInfo|MinSizeRel (default Debug)
  • CMAKE_<LANG>_CLANG_TIDY=clang-tidy (where <LANG> is C or CXX, for integrated static analysis, clang-tidy must be in PATH)
  • CMAKE_<LANG>_CPPLINT=cpplint (where <LANG> is C or CXX, for integrated style checking, cpplint must be in PATH)
  • CMAKE_<LANG>_INCLUDE_WHAT_YOU_USE (where <LANG> is C or CXX, include-what-you-use must be in PATH)
  • CODE_COVERAGE=ON|OFF (default OFF) - compute Code Coverage using gcov and lcov. See note on Code Coverage below.
  • CMAKE_INSTALL_PREFIX - set to the desired installation location

Build

To build, run the following from the build directory:

cmake --build .

Test

To run tests, run the following from the build directory:

ctest

Install

To install, run the following from the build directory:

cmake --install .

By default, the installation prefix will be an unversioned install/ directory at the project top-level. To customize this, set CMAKE_INSTALL_PREFIX when configuring.

Code Coverage

Code coverage can be computed by passing the CODE_COVERAGE=ON option during CMake configuration. Code coverage may NOT be enabled simultaneously with static analysis (namely, clang-tidy checks), because certain compilation options conflict between GCC and Clang. So if you want to compute code coverage, make a separate build directory just for coverage like so:

mkdir build_coverage && cd build_coverage
cmake -DCODE_COVERAGE=ON -DCMAKE_BUILD_TYPE=Debug -GNinja ..
cmake --build . --target ctest_coverage

Coverage should only be configured using the Debug build type. The commands above will configure coverage, and execute the test suite to determine runtime coverage. The HTML reports will be output to build_coverage/ctest_coverage and can be viewed by opening index.html in a web browser.

Customizing clang-tidy Checks

The top-level file .clang-tidy configures the rules that are checked or skipped by clang-tidy. A default set of rules is provided. Customize them by reviewing the available Clang-Tidy Checks and setting the appropriate options in the .clang-tidy file following the instructions.

Visual Studio Code IDE Support

Included in this project are configurations for editing, building and debugging using Microsoft VS Code.

At a minimum, you should install the following VS Code Extensions in order for the configure/build/debug commands to work seemlessly:

Note on Windows WSL

If you've cloned the project in WSL, open the project in VS Code from the WSL terminal by running the following from the project top-level:

code .

The first time this is run, VS Code server will install in the WSL environment, which will take some time on the initial launch. Subsequent launches will be faster.

Select Kit

Select the "kit" to be used, which defines the compilation toolchain for your target.

Open Command (Command+Shift+P), type CMake, select CMake: Scan for Kits and select one

Select Build Variant

Open Command (Command+Shift+P), type CMake, select CMake: Select Variant and select your desired build variant (Recommended: select Debug + Static Analysis and Style Checks + IWYU)

Note on Alpine Linux: IWYU is not available, so be sure to select one of the variants that says No IWYU.

Configure

Open Command (Command+Shift+P), type CMake, select CMake: Configure

Build

Open Command (Command+Shift+P), type CMake, select CMake: Build

Debug

Click the Run button in the VS Code sidebar (or Command+Shift+D) and run the Launch configuration.

Or, open Command (Command+Shift+P), type CMake, select CMake: Debug (make sure you first set a breakpoint somewhere, like main())

Run Tests

Open Command (Command+Shift+P), type CMake, select CMake: Run Tests

Other Commands

Open Command (Command+Shift+P), type CMake, select desired command

About

A C/C++ project template using modern CMake best practices.

License:Other


Languages

Language:C 54.3%Language:CMake 34.5%Language:Python 5.2%Language:C++ 4.4%Language:Dockerfile 1.6%