Installing Or Building With CMake#
Slint comes with a CMake integration that automates the compilation step of the .slint
markup language files and
offers a CMake target for convenient linkage.
Note: We recommend using the Ninja generator of CMake for the most efficient build and .slint
dependency tracking.
You can select the CMake Ninja backend by passing -GNinja
or setting the CMAKE_GENERATOR
environment variable to Ninja
.
Binary Packages#
We offer binary packages of Slint for use with C++. These work without any Rust development environment.
You can download one of our pre-built binaries for Linux or Windows on x86-64 architectures:
Click on the latest release
From “Assets” download either
slint-cpp-XXX-Linux-x86_64.tar.gz
for a Linux archive orslint-cpp-XXX-win64.exe
for a Windows installer. (“XXX” refers to the version of the latest release)Unpack the downloaded archive or run the installer.
After extracting the artifact or running the installer, you need to place the installation
directory into your CMAKE_PREFIX_PATH
by using the -DCMAKE_PREFIX_PATH=/path/to/installed/slint
argument in your cmake invocation. find_package(Slint)
will
then be able to find Slint from within a CMakeLists.txt
file.
At runtime you might also need to add the lib
sub-directory to the PATH
environment variable on Windows or the LD_LIBRARY_PATH
on Linux. This is
necessary to find the Slint libraries when trying to run your program.
In the next section you will learn how to use the installed library in your application
and how to work with .slint
UI files.
Building From Sources#
The recommended and most flexible way to use the C++ API is to build Slint from sources.
First you need to install the prerequisites:
Install Rust by following the Rust Getting Started Guide. If you already have Rust installed, make sure that it’s at least version 1.60 or newer. You can check which version you have installed by running
rustc --version
. Once this is done, you should have therustc
compiler and thecargo
build system installed in your path.cmake (3.21 or newer)
A C++ compiler that supports C++20 (e.g., MSVC 2019 16.6 on Windows)
You can include Slint into your CMake project using CMake’s
FetchContent
feature. Insert the following snippet into your CMakeLists.txt
to make CMake
download the latest released 1.x version, compile it, and make the CMake
integration available:
include(FetchContent)
FetchContent_Declare(
Slint
GIT_REPOSITORY https://github.com/slint-ui/slint.git
# `release/1` will auto-upgrade to the latest Slint >= 1.0.0 and < 2.0.0
# `release/1.0` will auto-upgrade to the latest Slint >= 1.0.0 and < 1.1.0
GIT_TAG release/1
SOURCE_SUBDIR api/cpp
)
FetchContent_MakeAvailable(Slint)
If you prefer to use Slint as an external CMake package, then you build Slint from source like a regular
CMake project, install it into a prefix directory of your choice and use find_package(Slint)
in your CMakeLists.txt
.
Resource Embedding#
By default, images or fonts that your Slint files reference are loaded from disk at run-time. This minimises build times, but requires that the directory structure with the files remains stable. If you want to build a program that runs anywhere, then you can configure the Slint compiler to embed such sources into the binary.
Set the SLINT_EMBED_RESOURCES
target property on your CMake target to one of the following values:
embed-files
: The raw files are embedded in the application binary.embed-for-software-renderer
: The files will be loaded by the Slint compiler, optimized for use with the software renderer and embedded in the application binary.as-absolute-path
: The paths of files are made absolute and will be used at run-time to load the resources from the file system. This is the default.
This target property is initialised from the global DEFAULT_SLINT_EMBED_RESOURCES
cache variable. Set it to configure the default for all CMake targets.
Features#
The Slint library supports a set of features, not all of them enabled by default.
You might want to adapt the set of enabled features to optimize your binary
size. For example you might want to support only the wayland stack on Linux.
Enable the backend-winit-wayland
feature while turning off the
backend-winit-x11
feature to do so.
Slint’s CMake configuration uses CMake options prefixed with SLINT_FEATURE_
to
expose Slint’s feature flags at compile time. To have a wayland-only stack with
the CMake setup you would for example use:
cmake -DSLINT_FEATURE_BACKEND_WINIT=OFF -DSLINT_FEATURE_BACKEND_WINIT_WAYLAND=ON ...
Alternatively, you can use cmake-gui
or ccmake
for a more interactive way
to discover and toggle features.
This works when compiling Slint as a package, using cmake --build
and
cmake --install
, or when including Slint using FetchContent
.
Back-Ends#
Slint needs a back-end that acts as liaison between Slint and the OS. Several back-ends can be built into the Slint library at the same time, but only one is used a run time.
Compile Time Back-End Selection#
By default Slint will include both the Qt and
winit back-ends – if both are detected at
compile time. You can enable or disable back-ends using the
SLINT_FEATURE_BACKEND_
features. For example, to exclude the winit back-end,
you would disable the SLINT_FEATURE_BACKEND_WINIT
option in your CMake
project configuration.
The winit back-end needs a renderer. SLINT_FEATURE_RENDERER_WINIT_FEMTOVG
is
the only stable renderer, the other ones are experimental. If you disable the
SLINT_FEATURE_BACKEND_WINIT
, you will also want to disable the renderer!
Run Time Back-End Selection#
It’s also possible to select any of the compiled in back-ends and renderer at
runtime, using the SLINT_BACKEND
environment variable.
SLINT_BACKEND=Qt
selects the Qt back-endSLINT_BACKEND=winit
selects the winit back-endSLINT_BACKEND=winit-femtovg
selects the winit back-end with the femtovg rendererSLINT_BACKEND=winit-skia
selects the winit back-end with the skia rendererSLINT_BACKEND=winit-software
selects the winit back-end with the software renderer
If the selected back-end or renderer isn’t available, the default will be used instead.
Cross-compiling#
It’s possible to cross-compile Slint to a different target architecture when building with CMake. You need to make sure your CMake setup is ready for cross-compilation, as documented in the upstream CMake documentation.
If you are building against a Yocto SDK, it is sufficient to source the SDK’s environment setup file.
Since Slint is implemented using the Rust programming language, you need to determine which Rust target matches the target architecture that you’re compiling for. Please consult the upstream Rust documentation to find the correct target name. Now you need to install the Rust toolchain:
rustup target add <target-name>
Then you’re ready to iconfigure your CMake project you need to add
-DRust_CARGO_TARGET=<target name>
to the CMake command line.
This ensures that the Slint library is built for the correct architecture.
For example if you are building against an embedded Linux Yocto SDK targeting an ARM64 board, the following commands show how to compile:
Install the Rust targe toolchain once:
rustup target add aarch64-unknown-linux-gnu
Set up the environment and build:
. /path/to/yocto/sdk/environment-setup-cortexa53-crypto-poky-linux
cd <PROJECT_ROOT>
mkdir build
cd build
cmake -DRust_CARGO_TARGET=aarch64-unknown-linux-gnu -DCMAKE_INSTALL_PREFIX=/slint/install/path ...
cmake --build .
cmake --install .