This is a collection of CMake utilities to include in a C++ project a set of useful development tools.
You can include it using Nix, CPM, Git submodule or system-wide installation. Nix is the recommended way as it manage the installation of the tools themselves (which are usually either not in the distro repos or outdated).
- Lizard
- IWYU
- Uncrustify
- Cppcheck
- ClangBuildAnalyzer
- Clang-Format
- Clang-Tidy
- CodeChecker (Clang-Static-Analyzer)
- LaTeX (works with XeLaTeX compiler, minted & tikz packages)
- Icon utils (based on CairoSVG and ImageMagick)
- LibFuzzer
- Sanitizers
- Lcov / Llvm-cov / Gcovr (works with Catch2, with SonarQube / SonarCloud integration)
- Doxygen (with Coverxygen and m.css integration)
You can select the dependencies you need by setting the corresponding with-<tool>
flag to true
, except for LaTeX due to the particularity explained here, you have to declare yourself your LaTeX dependency. XeLaTeX required by the CMakeUtils module is in the scheme-small
package and better.
You can also depend on a different set of tools between regular build and development shell as on the sample below.
flake.nix
(using either flake package output, either overlay output)
{
inputs = {
nixpkgs.url = "github:NixOS/nixpkgs";
flake-utils.url = "github:numtide/flake-utils";
cmake-utils = {
url = "github:conformism/cmake-utils";
inputs.nixpkgs.follows = "nixpkgs";
};
};
outputs = { self, nixpkgs, cmake-utils, flake-utils, ... }:
flake-utils.lib.eachDefaultSystem (system:
let
# Using flake package output
pkgs = nixpkgs.legacyPackages.${system};
# Using overlay
pkgs = nixpkgs.legacyPackages.${system}.extend cmake-utils.overlays.pkgs;
in {
devShell = pkgs.mkShell {
inputsFrom = [
self.defaultPackage
# Using flake package output
(cmake-utils.packages.${system}.cmake-utils-full.override {
with-sonar = false;
})
# Using overlay
(pkgs.cmake-utils-full.override {
with-sonar = false;
})
];
};
# Using flake package output
defaultPackage = pkgs.callPackage ./default.nix {
inherit (cmake-utils.packages.${system}) cmake-utils;
};
# Using overlay
defaultPackage = pkgs.callPackage ./default.nix {};
});
}
default.nix
{ stdenv
, texlive
, cmake-utils
}:
let
texlive-combined = texlive.combine {
inherit (texlive) scheme-small standalone pgfplots;
};
in stdenv.mkDerivation {
src = ./.;
nativeBuildInputs = [
cmake
cmake-utils
texlive-combined
];
}
CMakeLists.txt
include( CPM )
CPMAddPackage(
NAME cmake-utils
GITHUB_REPOSITORY conformism/cmake-utils
GIT_TAG main
DOWNLOAD_ONLY
)
set( CMakeUtils_DIR "${cmake-utils_SOURCE_DIR}" )
CMakeLists.txt
list( PREPEND CMAKE_MODULE_PATH
"${CMAKE_SOURCE_DIR}/<3rd_party>/cmake-utils"
)
- Install CMakeUtils
sudo make install # uninstall target also exist.
- Download m.css if you need it
cd /opt
sudo git clone https://github.com/mosra/m.css
CMakeLists.txt
find_package( CMakeUtils )
set( CMAKE_UTILS
CLANG_BUILD_ANALYZER
CLANG_FORMAT
CLANG_TIDY
CODECHECKER
COVERAGE
CPPCHECK
DOXYGEN
ICON
IWYU
LATEX
LIBFUZZER
LIZARD
SANITIZERS
UNCRUSTIFY
)
# Or all tools at once
set( CMAKE_UTILS * )
include( CMakeUtils )
For more informations about those options, take a look to the utils detail paragraph.
Option | Default | Description | Concerned utils |
---|---|---|---|
BUILD_ANALYZER |
OFF |
Enable Clang build statistics | CLANG_BUILD_ANALYZER |
STATIC_WERROR |
OFF |
Change warning into errors in the static analysis tools | CLANG_FORMAT CLANG_TIDY |
STATIC_ERROR |
OFF |
Do not tolerate errors in the static analysis tools | CLANG_FORMAT CLANG_TIDY CPPCHECK IWYU LIZARD UNCRUSTIFY |
CODECHECKER_REPORT |
OFF |
Tools that are supported as analyzers produce CodeChecker reports | CODECHECKER CPPCHECK LIBFUZZER |
COVERAGE |
OFF |
Enable coverage for the current build type, prefer to use the Coverage build type | COVERAGE |
SONAR |
OFF |
The coverage target will produce SonarQube reports instead of console / HTML | COVERAGE |
COVERAGE_ERROR |
ON |
Abort on unittest error | COVERAGE |
COVERAGE_GLOBAL_ONLY |
OFF |
When calling the 'coverage' target, do not show the dependent per target reports. The counterpart is the creation of intermediate targets | COVERAGE |
COVERAGE_GCOVR_VERBOSE |
OFF |
Print gcovr reports in terminal while generating sonarqube coverage reports | COVERAGE |
DOXYGEN_MCSS |
OFF |
Enable m.css Doxygen reports | DOXYGEN |
MCSS_VERSION |
|
M.CSS git tag | DOXYGEN |
LATEX_VERBOSE |
OFF |
Show LaTeX compiler output messages. | LATEX |
LIBFUZZER |
OFF |
Enable LibFuzzer | LIBFUZZER |
UNCRUSTIFY_DIFF |
ON |
Show diff of the suggested modifications while running uncrustify | UNCRUSTIFY |
SANITIZER |
|
Compile with a sanitizer. Options are: ASAN , AUBSAN , CFISAN , LSAN , MSAN , MWOSAN , TSAN , UBSAN |
SANITIZERS LIBFUZZER |
enable_clang_build_analyzer(
TARGET <target>
)
TARGET
: Target to analyse.
clang_format(
TARGET <target>
ADDITIONAL_FILES
<file1>
<...>
ARGS
<arg1>
<...>
)
TARGET
: Target to analyse. Every source file will be analysed.ADDITIONAL_FILES
: Specify other files to analyse, headers for instance.ARGS
: Specify command line arguments.
clang_tidy(
TARGET <target>
ADDITIONAL_FILES
<file1>
<...>
ARGS
<arg1>
<...>
)
TARGET
: Target to analyse. Every source file will be analysed.ADDITIONAL_FILES
: Specify other files to analyse, headers for instance.ARGS
: Specify command line arguments.
codechecker(
TARGET <target>
GLOBAL
NO_CTU
ADDITIONAL_OPTIONAL_REPORTS
<dir1>
<...>
SKIP
<arg1>
<...>
ARGS
<arg1>
<...>
)
TARGET
: Target to analyse. Will set codechecker target name in consequences.GLOBAL
: Create a global codechecker target instead of a per-target one, should be preferred to cover a whole project.NO_CTU
: Disable cross translation unit analysis.ADDITIONAL_OPTIONAL_REPORTS
: Specify other analysis reports, generated by tools supported by the report-converter program. Enable report export from those tools with theCODECHECKER_REPORT
option.SKIP
: Specify files to analyse regarding the codechecker skipfile syntax.ARGS
: Specify codechecker analyse command line arguments.
coverage(
TARGET_TO_RUN <target>
TARGETS_TO_COVER
<target1>
<...>
EXCLUDE_FROM_ALL
ARGS_GCOVR
<arg1>
<...>
ARGS_RUN
<arg1>
<...>
)
TARGET_TO_RUN
: Target to run, typically a unittest executable.TARGETS_TO_COVER
: Targets to produce coverage reports on, typically libraries tested by the unittest executable.EXCLUDE_FROM_ALL
: Exclude the coverage target from global reports produced by 'coverage' target.ARGS_GCOVR
: Arguments to pass to gcovr (used for sonarqube reports), especially useful to specify source coverage filters (-e
and-f
) as this is not automatically done as for lcov and llvm-cov.ARGS_RUN
: Arguments to pass to the unittest executable, especially useful to filter the executed tests during a development phase.
coverage_global()
cppcheck(
TARGET <target>
ADDITIONAL_FILES
<file1>
<...>
ARGS
<arg1>
<...>
)
TARGET
: Target to analyse. Every source file will be analysed.ADDITIONAL_FILES
: Specify other files to analyse, headers for instance.ARGS
: Specify command line arguments.
doxygen(
TARGET <target>
DOXYFILE <file>
MCSS_CONF <file>
EXCLUDE_FROM_ALL
TARGETS_TO_DOC
<target1>
<...>
ADDITIONAL_FILES
<file1>
<...>
ARGS_COVERXYGEN
<arg1>
<...>
)
TARGET
: Target name, better add_dox
suffix.TARGETS_TO_DOC
: Targets to run Doxygen on.ADDITIONAL_FILES
: Specify other files to analyse, headers for instance.DOXYFILE
: Doxyfile path. Following values are required:CLANG_DATABASE_PATH = @DOXYGEN_CLANG_DATABASE_PATH@
GENERATE_HTML = @DOXYGEN_GENERATE_HTML@
GENERATE_XML = @DOXYGEN_GENERATE_XML@
HTML_OUTPUT = @DOXYGEN_HTML_OUTPUT@
OUTPUT_DIRECTORY = @DOXYGEN_OUTPUT_DIRECTORY@
XML_OUTPUT = @DOXYGEN_XML_OUTPUT@
XML_PROGRAMLISTING = @DOXYGEN_XML_PROGRAMLISTING@
MCSS_CONF
: M.CSS configuration file path. Following values are required:DOXYFILE = @MCSS_DOXYFILE@
@MCSS_DOXYGEN_COVERAGE_INDEX@
may be used inLINKS_NAVBAR1
orLINKS_NAVBAR2
to add a link to Doxygen coverage Lcov report.
ARGS_COVERXYGEN
: Specify Coverxygen command line arguments.--src-dir
is required.EXCLUDE_FROM_ALL
: Exclude the Doxygen target from global call todox
target.
svg_to_png( <name>
DESTINATION <destination_dir>
SOURCE <source_dir>
SIZE <size>
REGISTER_TO <var>
)
DESTINATION
: Directory where output png file is created. Default is :${CMAKE_CURRENT_BINARY_DIR}
.SOURCE
: Directory where input svg file is located. Default is :${CMAKE_CURRENT_SOURCE_DIR}
.SIZE
: Desired output size in pixels. Both formats<size>
and<width>x<height>
are accepted.REGISTER_TO
: Append output png file to var. To the png file to be compiled, at least one target must DEPENDS on its output file. Avoid making more than one target DEPENDS on one output pdf file, instead create dependencies between other dependent targets and the one that wraps the png file compilation.
png_to_ico( <name>
DESTINATION <destination_dir>
INPUT
<file1>
<...>
REGISTER_TO <var>
)
DESTINATION
: Directory where output ico file is created. Default is :${CMAKE_CURRENT_BINARY_DIR}
.INPUT
: Input png files to combine in an ico file, typically multiple sizes of of the same image. Inclusion order seems to matter when using the file as an icon.REGISTER_TO
: Append output ico file to var. To the ico file to be compiled, at least one target must DEPENDS on its output file. Avoid making more than one target DEPENDS on one output pdf file, instead create dependencies between other dependent targets and the one that wraps the ico file compilation.
iwyu(
TARGET <target>
ADDITIONAL_FILES
<file1>
<...>
ARGS
<arg1>
<...>
)
TARGET
: Target to analyse. Every source file will be analysed.ADDITIONAL_FILES
: Specify other files to analyse, headers for instance.ARGS
: Specify command line arguments.
compile_latex_file( <name>
OUTPUT
<output2>
<...>
DESTINATION <destination_dir>
SOURCE <source_dir>
TEXINPUTS <texinputs>
SUBDIRS
<subdir1>
<...>
REGISTER_TO <var>
DEPENDS
<file1>
<...>
SHELL_ESCAPE
MINTED
)
OUTPUT
:${destination_dir}/${name}.pdf
is the default output file. You can add some others with this flag.DESTINATION
: Directory where output pdf file and byproduct files are created. Default is :${CMAKE_CURRENT_BINARY_DIR}
.SOURCE
Directory where main tex file is located. Default is :${CMAKE_CURRENT_SOURCE_DIR}
.TEXINPUTS
: Directories where other files (tex files, images, classes, etc.) are included from. Default contains current directory (at build time) and LaTeX system directories. Should be set as long as main tex file includes other files AND you don't execute the build command where the main tex file is located. Or if a tex file includes another file from elsewhere than the main tex file directory. Syntax is :dir1:dir2:...
SUBDIRS
: Directories where other tex files are located. Should be a relative path from aTEXINPUTS
location. The purpose of this flag is to create a build tree that correspond to the source tree. For example, if a tex file contains this\include{chapters/chapter2}
, addchapters
to subdirs.REGISTER_TO
: Append output pdf file to var. To the tex file to be compiled, at least one target mustDEPENDS
on its output file. Avoid making more than one targetDEPENDS
on one output pdf file, instead create dependencies between other dependent targets and the one that wraps the tex file compilation.DEPENDS
: By default a rebuild of the LaTeX document is triggered only if the main tex file is newer than the output file. This argument lets you trigger a rebuild on additional input files' timestamp.SHELL_ESCAPE
: Append--shell-escape
to the LaTeX compiler.MINTED
: Use it only if you use the LaTeX package Minted AND you overrideDESTINATION
. Defines\mintedoutputdir
to set the Minted package argumentoutputdir
. So you can include Minted this way :\usepackage[outputdir=\mintedoutputdir]{minted}
.
add_libfuzzer_target(
TARGET <target>
SOURCE <file>
TARGET_TO_FUZZ <target>
EXCLUDE_FROM_ALL
ARGS
<arg1>
<...>
)
TARGET
: Fuzzing target name.SOURCE
: Source file containing the libfuzzer entry point.TARGET_TO_FUZZ
: Target tested by the fuzzing target.EXCLUDE_FROM_ALL
: Exclude the fuzzing target from global call tofuzz
target.ARGS
: Arguments to pass to the fuzzing target run.
lizard(
TARGET <target>
ADDITIONAL_FILES
<file1>
<...>
ARGS
<arg1>
<...>
)
TARGET
: Target to analyse. Every source file will be analysed.ADDITIONAL_FILES
: Specify other files to analyse, headers for instance.ARGS
: Specify command line arguments.
enable_sanitizers(
TARGET <target>
)
TARGET
: Target to sanitize.
uncrustify(
TARGET <target>
ADDITIONAL_FILES
<file1>
<...>
ARGS
<arg1>
<...>
)
TARGET
: Target to analyse. Every source file will be analysed.ADDITIONAL_FILES
: Specify other files to analyse, headers for instance.ARGS
: Specify command line arguments.