I am trying to compile a project under MSYS2 and CLANG64 environment.
I have previously compiled dependencies in /usr/local.
$ ls /usr/local/include
boost compat-5.3.c cryptopp lauxlib.h libmongoc-1.0 lua.hpp luajit.h mongocxx yaml-cpp
bsoncxx compat-5.3.h gtest libbson-1.0 lua.h luaconf.h lualib.h tsl
$ ls /usr/local/lib
cmake libboost_filesystem-mt-s-x64.a libbson-static-1.0.a libmongoc-1.0.dll.a
libboost_atomic-mt-s-x64.a libboost_program_options-mt-s-x64.a libbsoncxx-static.a libmongoc-static-1.0.a
libboost_atomic-mt-x64.a libboost_regex-mt-s-x64.a libcryptopp.a libmongocxx-static.a
libboost_chrono-mt-s-x64.a libboost_system-mt-s-x64.a libgtest.a libyaml-cpp.a
libboost_container-mt-s-x64.a libboost_thread-mt-s-x64.a libgtest_main.a pkgconfig
libboost_context-mt-s-x64.a libbson-1.0.dll.a liblua-compat.a
But when I create the project, I explicitly set the location of binaries with interface libraries as I don't want to rely on the find mechanism that has hurt me badly in the past - linking to unintended, old system libraries.
project(test)
cmake_minimum_required( VERSION 3.0 )
add_library( cryptopp STATIC IMPORTED GLOBAL )
set_target_properties( cryptopp PROPERTIES
IMPORTED_LOCATION "/usr/local/lib/libcryptopp.a"
INTERFACE_INCLUDE_DIRECTORIES "/usr/local/include"
INTERFACE_COMPILE_DEFINITIONS "HAVE_CRYPTOPP"
)
add_executable( test test.cpp )
target_link_libraries( test cryptopp )
This works perfect under all Linux distros - Redhat, Ubuntu, etc but fails in MSYS2.
However when I run cmake, I get an error stating that /usr/local/include does not exist.
$ cmake ..
-- Building for: Ninja
-- The C compiler identification is Clang 14.0.4
-- The CXX compiler identification is Clang 14.0.4
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: G:/msys64/clang64/bin/cc.exe - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: G:/msys64/clang64/bin/c++.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
CMake Error in CMakeLists.txt:
Imported target "cryptopp" includes non-existent path
"/usr/local/include"
in its INTERFACE_INCLUDE_DIRECTORIES. Possible reasons include:
* The path was deleted, renamed, or moved to another location.
* An install or uninstall procedure did not complete successfully.
* The installation package was faulty and references files it does not
provide.
-- Generating done
CMake Generate step failed. Build files cannot be regenerated correctly.
I just cannot figure out why this is happening. Any clues?
Maybe it's a Windows path issue. Try replacing /usr/local with the output of cygpath -m /usr/local.
Related
cmake build failed with
CMAKE_AR-NOTFOUND cr libperfutils.a
When i checked build folder ar is not set.
x86_64-linux\gtest-native\1.7.0-r5\build\CMakeCache.txt
CMAKE_AR:FILEPATH=CMAKE_AR-NOTFOUND
//CXX compiler.
CMAKE_CXX_COMPILER:FILEPATH=/usr/bin/g++
How to set ar in cmake ?
Wheni see the cmake config below output . cmake automatically finding the g++
& why it not detecting ar
-- The CXX compiler identification is GNU 4.8.4
-- Check for working CXX compiler: g++
-- Check for working CXX compiler: g++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
When i check ar in machine manually its available.
/usr/bin/ar
CMake should have been able to detect /usr/bin/ar by itself, but you can manually override this, with the -D CMAKE_AR=/usr/bin/ar option.
Here's the list of useful variables:
https://gitlab.kitware.com/cmake/community/wikis/doc/cmake/Useful-Variables
cmake ar should set in cmake lib build ,
temperorly solved the issue by manally setting in my local Cmake file
set( CMAKE_AR "${AR}" CACHE FILEPATH "Archiver" )
I'm using cmake to compile a C++ project, and I want cmake to generate all the output files(metafiles like Makefile used to create binaries) in the build folder. I've checked all the answers in How do I make cmake output into a 'bin' dir?, none of them worked for me(suprisingly!). Files are generated in the root folder instead of in the build folder, what's wrong here? I guess I must have missed something.
Code Structure
➜ cmake-test tree .
.
├── CMakeLists.txt
└── hello.cpp
0 directories, 2 files
CMakeLists.txt
# Specify the minimum version for CMake
cmake_minimum_required(VERSION 3.11)
# Project's name
project(hello)
# Set the output folder where your program will be created
set(CMAKE_ARCHIVE_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/build)
set(CMAKE_LIBRARY_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/build)
set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)
# The following folder will be included
include_directories("${PROJECT_SOURCE_DIR}")
add_executable(hello ${PROJECT_SOURCE_DIR}/hello.cpp)
Build Commands and Outputs
➜ cmake-test cmake .
-- The C compiler identification is GNU 8.2.0
-- The CXX compiler identification is GNU 8.2.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
-- Generating done
-- Build files have been written to: /home/searene/CLionProjects/cmake-test
➜ cmake-test ls
bin CMakeCache.txt CMakeFiles cmake_install.cmake CMakeLists.txt hello.cpp Makefile
cmake version
➜ cmake-test cmake --version
cmake version 3.11.4
CMake suite maintained and supported by Kitware (kitware.com/cmake).
OS
Linux
The usual way to do this, rather than changing variables to set the path, is simply to create the output directory, change to it, and run cmake from there. So instead of cmake . you usually have cmake .. or similar.
I understand the initial impulse to say "But I expect my build system to write output somewhere else." But CMake is not usually used in the way you were initially expecting, and other people who run your CMake build won't expect what you were expecting, so it's probably best to just use the built-in, default behavior, which is to put the output wherever cmake was run.
Put another way: You are fighting against the tool. Don't do that.
Disclaimer: I recommend going with #john-zwinck's answer.
By default, cmake uses the current working directory as build directory and whatever path you provide as source directory. So the normal way of achieving your goal is
create the build directory (mkdir build)
go there (cd build)
call cmake with the source dir as argument (cmake path/to/source)
BUT there is another way, as far as I know not documented in the cmake docs and only kept for compatibility reasons or internal usage, that some people are using. The -B and -H flags
cmake -Hpath/to/source -Bpath/to/build
or even from the source dir
cmake . -Bbuild
Important: no space after -B.
CMake 3.19.1 (not sure how about older ones) has following option (from docs):
cmake [<options>] -S <path-to-source> -B <path-to-build>
Uses as the build tree and as the
source tree. The specified paths may be absolute or relative to the
current working directory. The source tree must contain a
CMakeLists.txt file. The build tree will be created automatically if
it does not already exist. For example:
cmake -S src -B build
I have bash scripts that generate c++ files and I would like to call them from an add_custom_command() in cmake. On windows, I'd like to call the bash script through wsl (windows subsytem for linux). cmake configures fine (finding wsl.exe), but on build it complains that:
'C:\Windows\System32\wsl.exe' is not recognized as an internal or external command,
Here is a simple cmake script to recreate. It succeeds in a linux environment (including wsl), but fails in the native windows environment.
cmake_minimum_required(VERSION 3.7)
project(WhyNotWSL)
set(source ${CMAKE_SOURCE_DIR}/Input/main.cpp)
set(target generated.cpp )
if (WIN32)
find_program(WSL wsl)
message("WSL is ${WSL}")
set (command ${WSL} cp ${source} ${target})
else()
set (command cp ${source} ${target})
endif()
message("command is ${command}")
add_custom_target( ${target} )
add_custom_command(
OUTPUT ${target}
COMMAND ${command}
DEPENDS ${source}
COMMENT "Generated ${target}"
)
add_custom_target(p ALL
DEPENDS ${target}
)
add_executable(hello ${target})
The following output is from cmake configure:
-- Building for: Visual Studio 16 2019
-- Selecting Windows SDK version 10.0.18362.0 to target Windows 10.0.18363.
-- The C compiler identification is MSVC 19.27.29110.0
-- The CXX compiler identification is MSVC 19.27.29110.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Professional/VC/Tools/MSVC/14.27.29110/
bin/Hostx64/x64/cl.exe - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: C:/Program Files (x86)/Microsoft Visual Studio/2019/Professional/VC/Tools/MSVC/14.27.2911
0/bin/Hostx64/x64/cl.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
**WSL is C:/Windows/System32/wsl.exe**
**command is C:/Windows/System32/wsl.exe;cp;D:/Development/CMakeWSLTest/Input/main.cpp;generated.cpp**
-- Configuring done
-- Generating done
-- Build files have been written to: D:/Development/CMakeWSLTest/BUILD
And the following is the output from cmake build
Microsoft (R) Build Engine version 16.7.0+b89cb5fde for .NET Framework
Copyright (C) Microsoft Corporation. All rights reserved.
Checking Build System
Building Custom Rule D:/Development/CMakeWSLTest/CMakeLists.txt
Building Custom Rule D:/Development/CMakeWSLTest/CMakeLists.txt
Generated generated.cpp
'C:\Windows\System32\wsl.exe' is not recognized as an internal or external command,
operable program or batch file.
For the record, the following works from a DOS shell:
>C:\Windows\System32\wsl.exe cp ../Input/main.cpp generated.cpp
Try giving HINTS to find_program() and point to C:\Windows\Sysnative
Something like this:
if (WIN32)
find_program(WSL wsl HINTS C:/Windows/Sysnative)
message("WSL is ${WSL}")
set (command ${WSL} cp ${source} ${target})
else()
set (command cp ${source} ${target})
endif()
Rationale here: https://superuser.com/a/1528297/2201 but basicly, this could be caused by your cmake buil. find_program() finds executable with absolute path that then should be different when it actually gets called during the build, it should actually point the binary from another location due to inherited build type of parent process of your build.
Maybe..
I would like to use CMAKE to compile a special piece of code in C99 with language extensions. Therefore I have to use a "custom" compiler and linker /your/path/to/compiler and /your/path/to/linker. How can I define the compiler and the linker command used by CMAKE?
EDIT:
I tried to define the compiler and linker as suggested by Equod:
set(CMAKE_C_COMPILER /your/path/to/compiler)
set(CMAKE_CUSTOM_LINKER /your/path/to/linker)
set(CMAKE_C_LINK_EXECUTABLE
"<CMAKE_CUSTOM_LINKER> <FLAGS> <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <OBJECTS> -o <TARGET> <LINK_LIBRARIES>")
But CMAKE is still not taking it:
-- Building for: Visual Studio 15 2017
-- Selecting Windows SDK version 10.0.14393.0 to target Windows 10.0.18363.
-- The C compiler identification is MSVC 19.16.27040.0
-- The CXX compiler identification is MSVC 19.16.27040.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: C:/Programms/VisualStudio/2017/Professional/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x86/cl.exe - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: C:/Programms/VisualStudio/2017/Professional/VC/Tools/MSVC/14.16.27023/bin/Hostx86/x86/cl.exe - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
-- Generating done
-- Build files have been written to: C:/your/path/to/source/build
What I forgot to mention before, I am working on a Windows machine and the executable of th compiler and linker is in the PATH.
EDIT:
Here is my CMakeLists.txt
cmake_minimum_required(VERSION 3.10)
project(MyProject
VERSION 1.0
DESCRIPTION "This is MyProject"
LANGUAGES C
)
set(CMAKE_C_COMPILER my_compiler)
set(CMAKE_C_LINK_EXECUTABLE my_linker)
configure_file(include/myproject_config.h.in include/myproject_config.h)
set(HEADER_FILES include/main.h include/somefunc.h)
set(SOURCE_FILES src/main.c src/somefunc.c)
add_executable(MyProject ${HEADER_FILES} ${SOURCE_FILES})
target_include_directories(MyProject PUBLIC "${PROJECT_BINARY_DIR}" )
target_include_directories(MyProject PUBLIC "../include" )
P.S.: my_compiler and my_linker are in PATH of cmd and PowerShell.
EDIT:
I installed MinGW now. I have make in my PATH as well. I updated the CMakeLists.txt file above. The make command tries to compile the code now with:
my_compiler #CMakeFiles/MyProject.dir/includes_C.rsp -o CMakeFiles\MyProject.dir\src\main.c.obj -c "C:\mypath\main.c"
But this is not working because I need a command like:
my_compiler -I="../include" "C:\mypath\main.c"
What CMake commands do I need to configure such a behavior?
CMake as default C compiler uses CC environment variables. You can also specify a different compiler and linker setting CMake variables:
Compiler:
set(CMAKE_C_COMPILER /your/path/to/compiler)
Linker:
set(CMAKE_CUSTOM_LINKER /your/path/to/linker)
set(CMAKE_C_LINK_EXECUTABLE
"<CMAKE_CUSTOM_LINKER> <FLAGS> <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <OBJECTS> -o <TARGET> <LINK_LIBRARIES>")
To better understand why linker settings are done in two steps I'd suggest taking a look at how those variables are managed internally by CMake:
CMake on github
I am attempting to build the ev3dev C++ bindings (see https://github.com/ddemidov/ev3dev-lang-cpp) for my EV3. But CMake isn't using my cross-compiler toolchain. I've successfully installed the code sourcery toolchain, built a simple hello world app, and run it on my brick. That works just fine. The trouble is that CMake uses my host's g++ compiler instead of the toolchain. I'm familiar with linux and cross-compilers, but I'm a total CMake noob. So I suspect CMake is where I'm going wrong.
I've added the following lines into CMakeLists.txt:
set(CMAKE_C_COMPILER "/cygdrive/c/Users/me/MentorGraphics/Sourcery_CodeBench_Lite_for_ARM_GNU_Linux/bin/arm-none-linux-gnueabi-gcc.exe")
set(CMAKE_CXX_COMPILER "/cygdrive/c/Users/me/MentorGraphics/Sourcery_CodeBench_Lite_for_ARM_GNU_Linux/bin/arm-none-linux-gnueabi-g++.exe")
Here's the host gcc compiler:
$ gcc --version
gcc (GCC) 5.4.0
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
And the target toolchain gcc compiler:
$ /cygdrive/c/Users/me/MentorGraphics/Sourcery_CodeBench_Lite_for_ARM_GNU_Linux/bin/arm-none-linux-gnueabi-gcc.exe --version
arm-none-linux-gnueabi-gcc.exe (Sourcery CodeBench Lite 2014.05-29) 4.8.3 20140320 (prerelease)
Copyright (C) 2013 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
The compilers are different versions, so they are really easy to distinguish.
Now here's the terminal session for the cmake build:
$ mkdir build
$ cd build
$ cmake ../ -DEV3DEV_PLATFORM=EV3
-- No build type selected, default to RelWithDebInfo
-- The C compiler identification is GNU 5.4.0
-- The CXX compiler identification is GNU 5.4.0
CMake Warning at /usr/share/cmake-3.6.2/Modules/Platform/CYGWIN.cmake:15 (message):
CMake no longer defines WIN32 on Cygwin!
(1) If you are just trying to build this project, ignore this warning or
quiet it by setting CMAKE_LEGACY_CYGWIN_WIN32=0 in your environment or in
the CMake cache. If later configuration or build errors occur then this
project may have been written under the assumption that Cygwin is WIN32.
In that case, set CMAKE_LEGACY_CYGWIN_WIN32=1 instead.
(2) If you are developing this project, add the line
set(CMAKE_LEGACY_CYGWIN_WIN32 0) # Remove when CMake >= 2.8.4 is required
at the top of your top-level CMakeLists.txt file or set the minimum
required version of CMake to 2.8.4 or higher. Then teach your project to
build on Cygwin without WIN32.
Call Stack (most recent call first):
/usr/share/cmake-3.6.2/Modules/CMakeSystemSpecificInformation.cmake:36 (include)
CMakeLists.txt:7 (project)
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++.exe
-- Check for working CXX compiler: /usr/bin/c++.exe -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
-- Generating done
-- Build files have been written to: /cygdrive/c/Users/me/Documents/AppProjects/ev3dev-lang-cpp/build
$
CMake is clearly using the cygwin host gcc/g++ compiler even though I've attempted to configure it to use the toolchain. Where am I going wrong?
EDIT: This is not a duplicate of cmake: problems specifying the compiler (2). I did actually use google, and the cross-compiling instructions on the CMake wiki (https://cmake.org/Wiki/CMake_Cross_Compiling) explicitly dictate using SET(). The problem turned out to be that CMake requires the compiler to be set before PROJECT(). I just moved the SET() lines up above PROJECT() and then CMake found the compiler.