cunit-manager

c-unit

CUnit is the answer to c/c++ dependency management in the CMake world. It comes as an amalgamation of other attempts including awesome projects like mason, hunter & conan.

We feel that each of those projects have respective shortcomings and based on all of them, we have designed c-unit to be best in breed.

note c-unit has a strong emphasis on creating statically linked binary distributions, therefore most default recipes are for static compilation. You are more than welcome to create shared lib recipes and publish them.

Furthermore, you can issue global cmake override flags converting all dependencies to shared libs

install

cu is implemented in JS for flexibility, so installation anywhere is dead simple.

npm i -g cunit-manager

Repos and Recipes

cu works with recipes in repositories. The global default repository is here: github.com/cunit/cunits

note you can add your own repos public or private both locally and git based as follows

# GITHUB EXAMPLE PUBLIC OR PRIVATE 
cunit repo add https://github.com/myorg/my-cunits.git
# OR A LOCAL DIR 
cunit repo add file:///var/cunits-local-on-disk

A recipe is super simple

name: opencv
repository: 
  url: https://github.com/opencv/opencv.git
 
cmake:
  flags:
    CMAKE_BUILD_TYPE: Release
    BUILD_SHARED_LIBS: OFF
    BUILD_JPEG: OFF
    BUILD_JASPER: OFF
    BUILD_PNG: ON
    BUILD_ZLIB: ON
    BUILD_IPP_IW: OFF
    BUILD_ITT: OFF
    BUILD_JAVA: OFF
    BUILD_PROTOBUF: OFF
    WITH_PROTOBUF: OFF
    WITH_CAROTENE: OFF
    WITH_CUBLAS: OFF
    WITH_CUDA: OFF
    WITH_CUFFT: OFF
    WITH_FFMPEG: OFF
    WITH_GPHOTO2: OFF
    WITH_GSTREAMER: OFF
    WITH_GTK: OFF
    WITH_ITT: OFF
    WITH_IPP: OFF
    WITH_JASPER: OFF
    WITH_LAPACK: OFF
    WITH_MATLAB: OFF
    WITH_NVCUVID: OFF
    WITH_OPENCL: OFF
    WITH_OPENCLAMDBLAS: OFF
    WITH_OPENCLAMDFFT: OFF
    WITH_OPENEXR: OFF
    WITH_PTHREADS_PF: OFF
    WITH_V4L: OFF
    WITH_WEBP: OFF
  findTemplate: cmake/FindOpenCV.cmake.hbs
 
dependencies:
  libtiff: Release-v4-0-9
  zlib: v1.2.11
  libpng: v1.6.33
  libjpeg: 8.4.0

Quickstart

It takes a few simple steps.

• Create a cunit.yml file in the root of your project

name: cunit-example
 
dependencies:
  protobuf: 3.1.0
  opencv: 3.4.1
 
tools:
    protobuf: 3.1.0

• Then run configure

cunit project configure

• In your root CMakeLists.txt BEFORE your project declaration add something like the following:

cmake_minimum_required(VERSION 3.10)
 
# INSERT THIS LINE
include(${CMAKE_CURRENT_LIST_DIR}/.cunit/cunit.cmake)
 
project(cunit_example)

Cross-Compiling "Toolchains"

Create your standard cmake toolchain file and use it as follows:

name: cunit-example
profiles: [Debug,Release]
 
toolchains:
  "aarch64-linux-gnu": cmake/aarch64.cmake 
  # file would be at this relative location from the project root 
 
dependencies:
  protobuf: 3.1.0
  opencv: 3.4.1
 

In order to use with non-cmake dependencies and scripts add the following to the top of your toolchain file:

include(${CMAKE_CURRENT_LIST_DIR}/.cunit/cunit.toolchain.cmake)

and add the following to the bottom of your toolchain file

cunit_toolchain_export()