I wrote a CMakeLists.txt for a project in C++, which uses OpenCV libraries. When I try to create the project using cmake, I get the next configuration problem:
CMake Error at CMakeLists.txt:15 (find_package):
Could not find module FindOpenCV.cmake or a configuration file for package
OpenCV.
Adjust CMAKE_MODULE_PATH to find FindOpenCV.cmake or set OpenCV_DIR to the
directory containing a CMake configuration file for OpenCV. The file will
have one of the following names:
OpenCVConfig.cmake
opencv-config.cmake
The fact is that I have an environment variable for the path which I use in Visual Studio with no problems. If I don't include OpenCV, then I can configure and generate with no problem, but I need to solve the problem. I don't understand why cmake cannot find the OpenCV path or how to fix it.
I also used the recommendations mentioned in this link:
FindOpenCV.cmake
Does anybody had this problem too?
The error you're seeing is that CMake cannot find a FindOpenCV.cmake file, because cmake doesn't include one out of the box. Therefore you need to find one and put it where cmake can find it:
You can find a good start here. If you're feeling adventurous you can also write your own.
Then add it somewhere in your project and adjust CMAKE_MODULE_PATH so that cmake can find it.
e.g., if you have
CMakeLists.txt
cmake-modules/FindOpenCV.cmake
Then you should do a
set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} ${CMAKE_CURRENT_SOURCE_DIR}/cmake-modules)
In your CMakeLists.txt file before you do a find_package(OpenCV)
If you are on Linux, you just need to fill the OpenCV_DIR variable with the path of opencv (containing the OpenCVConfig.cmake file)
export OpenCV_DIR=<path_of_opencv>
apt-get install libopencv-dev
export OpenCV_DIR=/usr/share/OpenCV
the header of cpp file should contain:
#include
#include "opencv2/highgui/highgui.hpp"
#include
#include
not original cv.h
find / -name "OpenCVConfig.cmake"
export OpenCV_DIR=/path/found/above
I had this exact same problem. I fixed it by adding the following line to my FindOpenCV.cmake file. Put it anywhere at the top before the rest of the code.
set (OpenCV_DIR /home/cmake/opencv/compiled) #change the path to match your complied directory of opencv
Basically you are telling FindOpenCV.cmake where to find opencv files assuming the other compilation can find the FindOpenCV.cmake
I faced the same error. In my case this "OpenCVConfig.cmake" file is located in /usr/local/share/OpenCV. In CMakeLists.txt add the line
set(OpenCV_DIR /usr/local/share/OpenCV)
as suggested by the error message.
if you are on windows, you can add opencv path to OpenCV_DIR yourself.
(OpenCV_DIR is in the red region)
the path is like "D:/opencv244/build".
you can find file "OpenCVConfig.cmake" under the path.
Another possibility is to denote where you can find OpenCV_DIR in the CMakeLists.txt file. For example, the following cmake scripts work for me:
cmake_minimum_required(VERSION 2.8)
project(performance_test)
set(OpenCV_STATIC ON)
set(OpenCV_CUDA OFF)
set(OpenCV_DIR "${CMAKE_SOURCE_DIR}/../install")
find_package(OpenCV REQUIRED)
include_directories(${OpenCV_INCLUDE_DIRS})
link_directories(${OpenCV_LIB_DIR})
file(GLOB my_source_files ./src/*)
add_executable( performance_test ${my_source_files})
target_link_libraries(performance_test ${OpenCV_LIBS})
Just to remind that you should set OpenCV_STATIC and OpenCV_CUDA as well before you invoke OpenCVConfig.cmake. In my case the built library is static library that does not use CUDA.
On my Fedora machine, when I typed "make" I got an error saying it could not find "cv.h". I fixed this by modifying my "OpenCVConfig.cmake" file.
Before:
SET(OpenCV_INCLUDE_DIRS "${OpenCV_INSTALL_PATH}/include/opencv;${OpenCV_INSTALL_PATH}/include")
SET(OpenCV_LIB_DIR "${OpenCV_INSTALL_PATH}/lib64")
After:
SET(OpenCV_INCLUDE_DIRS "/usr/include/opencv;/usr/include/opencv2")
SET(OpenCV_LIB_DIR "/usr/lib64")
I am using Windows and get the same error message. I find another problem which is relevant.
I defined OpenCV_DIR in my path at the end of the line. However when I typed "path" in the command line, my OpenCV_DIR was not shown. I found because Windows probably has a limit on how long the path can be, it cut my OpenCV_DIR to be only part of what I defined. So I removed some other part of the path, now it works.
I had the same error, I use windows. I add "C:\opencv\build" (opencv folder) to path at the control pannel.
So, That's Ok!!
For me (on Ubuntu), I just run:
sudo apt-get install libopencv-dev
Followed #hugh-pearse 's and #leszek-hanusz 's answers, with a little tweak. I had installed opencv from ubuntu 12.10 repository (libopencv-)* and had the same problem. Couldn't solve it with export OpenCV_DIR=/usr/share/OpenCV/ (since my OpenCVConfig.cmake whas there). It was solved when I also changed some lines on the OpenCVConfig.cmake file:
# ======================================================
# Include directories to add to the user project:
# ======================================================
# Provide the include directories to the caller
#SET(OpenCV_INCLUDE_DIRS "${OpenCV_INSTALL_PATH}/include/opencv;${OpenCV_INSTALL_PATH}/include")
SET(OpenCV_INCLUDE_DIRS "/usr/include/opencv;/usr/include/opencv2")
INCLUDE_DIRECTORIES(${OpenCV_INCLUDE_DIRS})
# ======================================================
# Link directories to add to the user project:
# ======================================================
# Provide the libs directory anyway, it may be needed in some cases.
#SET(OpenCV_LIB_DIR "${OpenCV_INSTALL_PATH}/lib")
SET(OpenCV_LIB_DIR "/usr/lib")
LINK_DIRECTORIES(${OpenCV_LIB_DIR})
And that worked on my Ubuntu 12.10. Remember to add the target_link_libraries(yourprojectname ${OpenCV_LIBS}) in your CMakeLists.txt.
When you install the libraries in the c drive (windows). the CMakeLists.txt shoud be looking like below:
cmake_minimum_required(VERSION 3.0.0)
project(test_opencv VERSION 0.1.0)
include(CTest)
enable_testing()
find_package(OpenCV REQUIRED)
include_directories(${OpenCV_INCLUDE_DIRS})
add_executable(test_opencv main.cpp)
target_link_libraries(test_opencv ${OPENCV_LIBS})
set(CPACK_PROJECT_NAME ${PROJECT_NAME})
set(CPACK_PROJECT_VERSION ${PROJECT_VERSION})
include(CPack)
finding the package and include directories
when building the project in VS code. Run the visual studio code with admin rights as the OpenCV is installed inside C drive.
I searched for the best way to do this, but I was unable to find a clear answer.
Was anyone able to build a tensorflow-serving client using cmake?
I am having difficulties with generating CPP files from proto, since they are needed for prediction service. Those proto files also include proto files from tensorflow.
so far I have come up with this:
project(serving C CXX)
find_package(Protobuf REQUIRED)
file(GLOB_RECURSE proto_files RELATIVE ${serving_SOURCE_DIR}/tensorflow/
"${serving_SOURCE_DIR}/tensorflow/*.proto")
set(PROTOBUF_GENERATE_CPP_APPEND_PATH OFF)
include_directories(${PROTOBUF_INCLUDE_DIRS})
include_directories( ${CMAKE_CURRENT_BINARY_DIR} )
PROTOBUF_GENERATE_CPP(PROTO_SRCS PROTO_HDRS ${proto_files})
add_library(tf_protos ${PROTO_SRCS} ${PROTO_HDRS})
target_link_libraries(tf_protos PUBLIC ${PROTOBUF_LIBRARIES})
Cmake builds successfully but the make command gives me an error:
No rule to make target '../tensorflow/tools/proto_text/test.proto', needed by 'tensorflow/tools/proto_text/test.pb.cc'. Stop.
To overcome the problem of .proto includes not being found I used command
set(PROTOBUF_GENERATE_CPP_APPEND_PATH OFF)
which was explained here: https://groups.google.com/forum/#!topic/protobuf/eow2fNDUHvc
My current folder structure is
serving/
CmakeLists.txt
tensorflow/
tensorflow_serving/
apis/
Folder apis contains .proto files that are needed in the client implementation and they include .proto files from the folder tensorflow.
Is this even the right way to go?
Any help/advice would be much appreciated.
I was able to get it to work in the layout you have where the CMakeLists.txt file is placed in the same level as the serving repository here. You'll need to install Tensorflow too though (using tensorflow_cc).
However, you probably don't want to muck with a fork of the official tensorflow/serving repository so I went a step further and moved the CMakeLists.txt out so you can just submodule the official repository. I made an example here
The gist is that the protobuf CMake submodule expects proto files to be laid out in the same directory from which it's called. I made some modifications to the submodules to let us call it from a level above serving and to ensure it invokes the compiler with the include paths in the right order to support the nested structure of the proto files in serving/tensorflow_serving/apis/* (and placing it accordingly in the specified build directory)
Hopefully someone else with better knowhow can make this better!
These worked for me.
https://github.com/wardsng/inception_cmake
https://github.com/FloopCZ/tensorflow_cc
You can choose a private install directory instead of the default, e.g. /usr/local/...
cmake -DCMAKE_INSTALL_PREFIX= ..
I have been unable to figure out how to get CMake to find and set correct GLFW CMake constants when using CMake in VS2017. Any help will be greatly appreciated :).
I downloaded glfw3 through Microsoft's vcpkg tool. I have checked that files do physically exist in the directory that vcpkg puts them in (~\vcpkg\installed\x86-windows\include). I set up my CMakeSettings.json as per their docs here. I used that tutorial as a basis for getting GLFW to be set up correctly.
I then use find_package(glfw3 REQUIRED) to find the glfw3 library. This does not spit out any errors. Actually the CMakeLists.txt doesn't complain at all. It's at the compile stage where it complains.
After that I add glfw3 with target_link_libraries(exe ${GLFW3_LIBRARIES}) to the executable.
Then when I try and build a simple example (including the header file), the compilation fails because it cannot find GLFW/glfw3.h.
The error from MSVC:
fatal error C1083: Cannot open include file: 'GLFW/glfw3.h': No such file or directory
Here is my CMakeLists.txt for added reference:
cmake_minimum_required(VERSION 3.7)
project(learn-opengl)
find_package(glfw3 REQUIRED)
add_executable(learn-opengl main.cpp)
target_link_libraries(learn-opengl ${GLFW3_LIBRARIES})
GLFW3_LIBRARIES I got from glfw3Config.cmake by snooping around what vcpkg puts in the installed directory (~\vcpkg\installed\x86-windows\share\glfw3)
And just in case, the main.cpp:
#include <GLFW/glfw3.h>
int main()
{
return 0;
}
I have tried calling cmake from the command line as well, but to no avail that didn't work either.
Am I missing something? Did I perhaps misunderstand something in vcpkg documentation? I have really no idea what I am missing... :/ I should say, in addition, that I am fairly new to CMake as well.
Reformulating my previous comment as answer:
You should add the imported target glfw to your target_link_libraries command instead of ${GLFW3_LIBRARIES}.
The find_package(glfw3) generates an import target glfw. By making your target learn-opengl dependent on this imported target you specify both the library to link with and the include directories to use.
After hours of scouring the web and SO for a solution I'm at a standstill. Nothing has worked so far for me...
I'm on Windows, using CLion IDE which uses CMake. My goal is to correctly link SDL2 to my project and use it through #include "SDL.h" which is the correct way.
The format of my CMakeLists.txt file
Specifics regarding the directory where I should have put the MingW development library of SDL2
Any requirements regarding windows ENV variables that I might have to set.
My CMakeLists.txt looks like this:
cmake_minimum_required(VERSION 3.6)
project(sdl2Project)
set(CMAKE_CXX_STANDARD 11)
#This is where sdl2-config.cmake is located
set(CMAKE_PREFIX_PATH ${CMAKE_PREFIX_PATH} "C:/Users/MyUserName/CLibraries/SDL2-2.0.5/x86_64-w64-mingw32/lib/cmake/SDL2")
set(SOURCE_FILES main.cpp)
add_executable(sdl2Project ${SOURCE_FILES})
find_package(sdl2 REQUIRED)
target_include_directories(sdl2Project PUBLIC ${SDL2_INCLUDE_DIRS})
target_link_libraries(sdl2Project ${SDL2_LIBRARIES})
There is no FindSDL2.cmake file used.
The SDL2 library I downloaded from libsdl.org is located in:
C:/Users/MyUserName/CLibraries/SDL2-2.0.5/x86_64-w64-mingw32
I have no experience with CMake so I'm unable to truly understand where the problem stems from. What are the steps I need to take in order for it to find the library and link it correctly??
EDIT:
My Project structure is the following:
sdl2Project
cmake-build-debug
CMakeLists.txt
main.cpp
Looking in your FindSDL2.cmake, you need to provide an hint to CMake about where the library is installed. You could do this by setting an environment variable SDLDIR, but you shouldn't. General advice: you shouldn't use a CMake package that wasn't provided with the sources you're using.
Looking in sources of SDL2, root directory contains a file sdl2-config.cmake.in that should have been configured and installed in your install directory as sdl2-config.cmake: that's the package file you should use.
Am I right guessing the file C:/Users/MyUserName/CLibraries/SDL2-2.0.5/sdl2-config.cmake exists?
If yes, to allow CMake to find it, add your install directory to CMAKE_PREFIX_PATH, before calling find_package:
set(CMAKE_PREFIX_PATH
${CMAKE_PREFIX_PATH}
"C:/Users/MyUserName/CLibraries/SDL2-2.0.5"
)
find_package(sdl2 REQUIRED)
Note the use of "/" in the path instead of "\" which could be interpreted as escaping character. Quotes around the path are only necessary if the path contains whitespaces.
EDIT:
Moreover, you misused target_link_libraries with a wrong target: SDL2 which you don't build in your project, instead of sdl2Project.
You also used a wrong variable: SDL2_LIBRARY instead of SDL2_LIBRARIES; you can see the good variable name by looking in sdl2-config.cmake.
You may consider target_include_directories instead of include_directories, but again the variable name you used is wrong: SDL2_INCLUDE_DIR instead of SDL2_INCLUDE_DIRS.
Try:
target_include_directories(sdl2Project PUBLIC ${SDL2_INCLUDE_DIRS})
target_link_libraries(sdl2Project ${SDL2_LIBRARIES})
I have build libmemcached.a and copied it to /usr/local/lib on my mac and I have tried all the following options to link libmemcached.a and yet get compile time errors that libmemcached/memcached.h is not found.
link_libraries (${libmemcached})
include_directories(SYSTEM ${libmemcached})
link_directories("/usr/local/lib")
find_package(libmemcached.a REQUIRED)
link_libraries`enter code here`(libmemcached.a)
find_library(RESULT libmemcached.a PATHS /usr/local/lib)
target_link_libraries(dnsa_pcl libmemcached.a)
It is a simple -L -l using MakeFile. Not sure what needs to be done to make this work using CMakeLists. Any help is much appreciated.
you should use so called imported library, like it's described in the official documentation
The solution that worked for me is a bit weird. I had to both set the CMAKE_PREFIX_PATH to the directory that has the lib and also have include_directories() with the include folder of the source. For some reason, I was under the impression that the libmemcached.a had the header files as well.