Adding External Library to Zephyr - cmake

Context:
I am trying to add an external library which uses CMake to my Zephyr project. I have explored the modules documentation, but this does not seem a good fit as I am unable to modify the upstream library and would like to avoid forking.
To add the library, I am using FetchContent in my Cmake file. This is working successfully and I am able to download and build the files.
Problem:
When linking, I encounter a "Conflicting CPU architectures" error. After inspecting into compile_commands.json, I can see the libraries source code is not receiving the same CMAKE_ARGS as the other files, leading to the architecture mismatch.
I am looking for the suggested way of adding external libraries to a zephyr project, without using the module system.
FetchContent_Declare(
my-lib
GIT_REPOSITORY git#github.com:<HostRepo>/<repoName>.git
GIT_TAG v0.1.7
)
FetchContent_MakeAvailable(my-lib)
target_link_libraries(app PRIVATE my-lib)

I recommend doing the repo fetching with a toplevel west.yml manifest and including your lib with add_library() or the appropriate cmake function.

Related

How to build an external library downloaded with CMake FetchContent?

I have a program that depends on an external library (SDL for example). I want CMake to take care of that dependency for me, so I was looking into FetchContent. As far as I understand, this module simply downloads the source code so that information on the external library is available at configure time. For example:
include(FetchContent)
FetchContent_Declare(sdl
GIT_REPOSITORY <...>
)
FetchContent_GetProperties(sdl)
# sdl_POPULATED, sdl_SOURCE_DIR and sdl_BINARY_DIR are ready now
if(NOT sdl_POPULATED)
FetchContent_Populate(sdl)
endif()
At some point, however, I want to build that source code and link it to my main executable. How to do it the "modern CMake way"?
The recommended way to build external libraries from source as part of your build depends on what the external lib provides build-wise.
External lib builds with cmake
If the external lib builds with cmake then you could add the lib to your build via a add_subdirectory(${libname_SOURCE_DIR}) call. That way cmake will build the external lib as a subfolder ("subproject"). The CMakeLists.txt file of the external lib will have some add_library(ext_lib_name ...) statements in it. In order to then use the external lib in your targets (an application or library that depends on the external lib) you can simply call target_link_libraries(your_application <PRIVATE|PUBLIC|INTERFACE> ext_lib_name) https://cmake.org/cmake/help/latest/command/target_link_libraries.html
I had a quick look at this github repo https://github.com/rantoniello/sdl - (let me know if you are referring to another library) and it actually looks like it is building with cmake and that it allows clients to statically or dynamically link against it: https://github.com/rantoniello/sdl/blob/master/CMakeLists.txt#L1688-L1740
So, ideally your applicaiton should be able to do
add_executable(myapp ...)
target_link_libraries(myapp PRIVATE SDL2-static) // Statically link againt SDL2
Due to their CMakeLists.txt file the SDL2-static comes with properties (include directories, linker flags/commands) that will automatically propergate to myapp.
External lib does not build with cmake
If a external lib doesn't build with cmake then one can try to use add_custom_target https://cmake.org/cmake/help/latest/command/add_custom_target.html to build the library. Something along the lines of:
add_custom_target(myExternalTarget COMMAND <invoke the repo's build system>)
You'd then need to set the target properties that are important for clients yourself via the the proper cmake functions set_target_properties, target_include_directories ... A great writeup how to get started with these kinds of things: https://pabloariasal.github.io/2018/02/19/its-time-to-do-cmake-right/

How to include a library using CMAKE in a cross-platform way

I am trying to use the assimp library in a cross platform C++ project. I include the repo as a git submodule, so, effectively, if someone downloads my project they will also download the ASSIMP project.
After I go through the assimp build / CMAKE instructions and (on Linux) type make install and from then on in my project I can use:
target_link_libraries(${PROJECT_NAME} assimp)
However, there is no make install on Windows.
The only other way I have been able to include the library on Linux is to put (in my CmakeLists.txt file):
target_link_libraries(${PROJECT_NAME} ${CMAKE_SOURCE_DIR}/build/assimp/code/libassimp.so)
This is not cross platform as it hardcodes the name and location of the .so file which will not work on Windows.
How can I expose the library so that I can do something like target_link_libraries(${PROJECT_NAME} assimp) on all platforms?
My directory tree looks like:
- src
- include
- assimp
- bin
Where the assimp directory in the include directory is the git submodule
I think you're going about this the wrong way. You don't need to build assimp in a separate step from your project, and you don't need to make install to make it available.
There are a number of ways of handling third party dependencies in Cmake, since you've already chosen to submodule the assimp repository, we'll start there. Assuming assimp is located in the root of your repository in a directory called assimp/ this would be a barebones project including it:
cmake_minimum_required(VERSION 3.0)
project(Project myassimpproj)
# include your directories
include_directories(
${CMAKE_CURRENT_SOURCE_DIR}
)
# set any variables you might need to set for your app and assimp
set(BUILD_ASSIMP_TOOLS ON)
set(ASSIMP_BUILD_STATIC_LIB ON)
# add assimp source dir as a subdirectory, effectively making
# assimp's CMakeLists.txt part of your build
add_subdirectory(/path/to/assimp ${CMAKE_BINARY_DIR}/assimp)
add_executable(assimp_target main.cpp)
# be sure to link in assimp, use platform-agnostic syntax for the linker
target_link_libraries(assimp_target assimp)
There may be a better way of phrasing this using generator expressions syntax, but I haven't looked at assimp's CMakeLists.txt to know if it's supported (and this is a more generic way anyway.)
Not every project uses Cmake, so you may not be able to just add_subdirectory(). In those cases, you can effectively "fake" a user call to build them using their build commands on respective platforms. execute_process() runs a command at configure time add_custom_command() and add_custom_target() run commands at build time. You then create a fake target to make integration and cross your fingers they support Cmake someday.
You can also use the ExternalProject commands added to Cmake to create a custom target to drive download, update/patch, configure, build, install and test steps of an external project, but note that this solution and the next download the dependency rather than using the submodule'd source code.
Finally, I prefer to work with prebuilt dependencies, cuts down on build time, and they can be unit tested on their own outside of the project. Conan is an open source, decentralized and multi-platform package manager with very good support for C++ and almost transparent support for Cmake when used the right way. They have grown very stable in the last year. More information on how to use Conan with Cmake can be found here.

Encapsulating a JNI library inside a jar file

I am trying to develop a plugin for Fiji/ImageJ that relies on a native library (JNI).
The JNI library itself depends on libtiff and fftw. On OSX and Linux, I use the class NativeUtils and everything works fine.
On windows, I included binary versions of libtiff and fftw in the CMake package and managed to link the JNI library against those (either statically of dynamically). However, the resulting JNI module does not include libtiff or fftw and I obtain an error when I try to load the JNI library with NativeUtils.loadLibraryFromJar. This is also the case when I include the dependent .dll in the .jar since they are not extracted by NativeUtils.
Here are the relevant lines in CMakeLists.txt:
add_library(fftw STATIC IMPORTED GLOBAL)
set_target_properties(fftw PROPERTIES IMPORTED_LOCATION "${libdir}/libfftw3f-3.lib"
INTERFACE_INCLUDE_DIRECTORIES "${incdir}")
SWIG_ADD_LIBRARY(javainterf
TYPE MODULE
LANGUAGE java
SOURCES javainterf.i javainterf.c src1.c)
SWIG_LINK_LIBRARIES(javainterf libcode1 fftw)
add_jar(Foo
SOURCES ${CMAKE_CURRENT_BINARY_DIR}/java/foo1.java
INCLUDE_JARS java/resources/ij-1.51p.jar
VERSION ${JAR_VERSION})
add_dependencies(Foo javainterf)
add_custom_command(TARGET Foo POST_BUILD
COMMAND "${Java_JAR_EXECUTABLE}" -uf Foo-${JAR_VERSION}.jar
-C ${CMAKE_CURRENT_BINARY_DIR}/${CMAKE_CFG_INTDIR} ${JNI_LIB_NAME})
How would you make sure that all the dependencies are properly included in the jar and loaded?
You can't load library from inside JAR without extracting it in a first place.
Take a look here at full sample where native code is embedded inside JAR and extracted when needed.
https://github.com/mkowsiak/jnicookbook/tree/master/recipes/recipeNo031
Update
Well, in that case, when you need to pack more libs and you want to properly resolve locations, you need to play with runtime env a little bit.
Take a look here:
https://github.com/mkowsiak/jnicookbook/tree/master/recipes/recipeNo035
git clone https://github.com/mkowsiak/jnicookbook
cd jnicookbook/recipes/recipeNo035
make
make test
Have fun with JNI!

CMake package configuration files for upstream projects using Qt5 problems

I am working on a larger C++ library that is using CMake and depends on Qt.
We moved from Qt4 to Qt5 and now I encounter a problem when using our lib
in an upstream project. As a minimal working example demonstrating the problem please have a look at this repo:
https://github.com/philthiel/cmake_qt5_upstream
It contains two separate CMake projects:
MyLIB: a tiny library that uses QString from Qt5::Core.
It generates and installs package configuration files
MyLIBConfig.cmake, MyLIBConfigVersion.cmake, and MyLIBTargets.cmake
in order to be searchable by CMake find_package()
MyAPP: a tiny executable depending on MyLIB
The project uses find_package(MyLIB) and creates an executable that uses MyLIB
The problem is that CMake gives me the following error message when configuring the MyAPP project:
CMake Error at CMakeLists.txt:11 (add_executable):
Target "MyAPP" links to target "Qt5::Core" but the target was not found.
Perhaps a find_package() call is missing for an IMPORTED target, or an
ALIAS target is missing?
The reason for this behaviour is that in the automatically generated MyLIBTargets.cmake file the INTERFACE_LINK_LIBRARIES entry for Qt5 Core is the Qt5::Core symbol. Using Qt4, the absolute path to the Qt core lib was specified here.
Now, I simply can resolve this by using
find_package(Qt5Core 5.X REQUIRED)
in the MyAPP project.
However, I would like to know if this is the intended/generic way to go, i.e. requesting upstream projects of our lib to search for the required transitive Qt5 dependencies themselves, or if I probably misuse CMake here and need to change my configuration procedure?
The CMake docu on package file generation
https://cmake.org/cmake/help/v3.0/manual/cmake-packages.7.html
mentions that macros can be provided by the package configuration files to upstream. Maybe this would be the correct place to search for imported targets like Qt5 and break upstream configuration runs when these dependencies are not found?
Best,
Philipp
[edit of the edit] Full Source Example
You need to deliver a CMake config file for your project, and probably the ConfigFile should be generated via CMake itself (because you cannot know for shure where the user will install your software).
Tip, use the ECM cmake modules to ease the creation of that:
find_package(ECM REQUIRED NO_MODULE)
include(CMakePackageConfigHelpers)
ecm_setup_version(${PROJECT_VERSION}
VARIABLE_PREFIX ATCORE
VERSION_HEADER "${CMAKE_CURRENT_BINARY_DIR}/atcore_version.h"
PACKAGE_VERSION_FILE "${CMAKE_CURRENT_BINARY_DIR}/KF5AtCoreConfigVersion.cmake"
SOVERSION 1
)
configure_package_config_file("${CMAKE_CURRENT_SOURCE_DIR}/KF5AtCoreConfig.cmake.in"
"${CMAKE_CURRENT_BINARY_DIR}/KF5AtCoreConfig.cmake"
INSTALL_DESTINATION ${CMAKECONFIG_INSTALL_DIR}
)
and the KF5AtCoreConfig.cmake.in:
#PACKAGE_INIT#
find_dependency(Qt5Widgets "#REQUIRED_QT_VERSION#")
find_dependency(Qt5SerialPort "#REQUIRED_QT_VERSION#")
find_dependency(KF5Solid "#KF5_DEP_VERSION#")
include("${CMAKE_CURRENT_LIST_DIR}/KF5AtCoreTargets.cmake")
This will generate the correct FindYourSortware.cmake with all your dependencies.
[edit] Better explanation on what's going on.
If you are providing a library that will use Qt, and that would also need to find the Qt5 library before compilling the user's source, you need to provide yourself a FindYourLibrary.cmake code, that would call
find_package(Qt5 REQUIRED COMPONENTS Core Gui Widgets Whatever)
Now, if it's your executable that needs to be linked, use the Components instead of the way you are doing it now.
find_package(Qt5 REQUIRED COMPONENTS Core)
then you link your library with
target_link_libraries(YourTarget Qt5::Core)

Linking Shared library (which has dependency on other shared library) in CMake

I am trying to build an executable which links to a shared library (named 'caffe'). The shared library is dependent on another shared library (named 'cblas'). When I try to link to caffe in my CMake file it shows the following error:
libcblas.so.3, needed by libcaffe.so, not found (try using -rpath or
-rpath-link)
I am using the following statements in my CMakeLists.txt:
link_directories(${BINARIES}/lib)
add_library(CAFFE_LIBRARY SHARED IMPORTED)
set_target_properties(CAFFE_LIBRARY PROPERTIES IMPORTED_LOCATION ${BINARIES}/lib/libcaffe.so)
target_link_libraries(${PROJECT_NAME} CAFFE_LIBRARY)
Both 'cblas' and 'caffe' libraries are present in ${BINARIES}/lib folder.
Do I need to add cblas.so to target_link_libraries also? Also, i am not building caffe.so so building it via CMake and keeping it as a dependency is not an option
Is there any other feasible solution for the same problem where dependency tree of shared library needs to be resolved while linking?
Browsing through the library's GitHub tree, it seems to me that it provides a package config file. Therefore, if you have installed it in the normal way, you should be able to find it as a package, instead of defining the imported target yourself:
find_package(Caffe)
include_directories(${Caffe_INCLUDE_DIRS})
add_definitions(${Caffe_DEFINITIONS}) # ex. -DCPU_ONLY
add_executable(caffeinated_application main.cpp)
target_link_libraries(caffeinated_application ${Caffe_LIBRARIES})
The example above comes from the Caffe documentation on the topic.