Clang and MSVC already supports Modules TS from unfinished C++20 standard.
Can I build my modules based project with CMake or other build system and how?
I tried build2, it supports modules and it works very well, but i have a question about it's dependency management (UPD: question is closed).
CMake currently does not support C++20 modules.
See also the relevant issue in the CMake issue tracker. Note that supporting modules requires far more support from the build system than inserting a new compiler option. It fundamentally changes how dependencies between source files have to be handled during the build: In a pre-modules world all cpp source files can be built independently in any order. With modules that is no longer true, which has implications not only for CMake itself, but also for the downstream build system.
Take a look at the CMake Fortran modules paper for the gory details. From a build system's point of view, Fortran's modules behave very similar to the C++20 modules.
Update: CMake 3.20 introduces experimental support for Modules with the Ninja Generator (and only for Ninja). Details can be found in the respective pull request. At this stage, this feature is still highly experimental and not intended for production use. If you intend to play around with this anyway, you really should be reading both the Fortran modules paper and the dependency format paper to understand what you're getting into.
This works on Linux Manjaro (same as Arch), but should work on any Unix OS. Of course, you need to build with new clang (tested with clang-10).
helloworld.cpp:
export module helloworld;
import <cstdio>;
export void hello() { puts("Hello world!"); }
main.cpp:
import helloworld; // import declaration
int main() {
hello();
}
CMakeLists.txt:
cmake_minimum_required(VERSION 3.16)
project(main)
set(CMAKE_CXX_STANDARD 20)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_EXTENSIONS OFF)
set(PREBUILT_MODULE_PATH ${CMAKE_BINARY_DIR}/modules)
function(add_module name)
file(MAKE_DIRECTORY ${PREBUILT_MODULE_PATH})
add_custom_target(${name}.pcm
COMMAND
${CMAKE_CXX_COMPILER}
-std=c++20
-stdlib=libc++
-fmodules
-c
${CMAKE_CURRENT_SOURCE_DIR}/${ARGN}
-Xclang -emit-module-interface
-o ${PREBUILT_MODULE_PATH}/${name}.pcm
)
endfunction()
add_compile_options(-fmodules)
add_compile_options(-stdlib=libc++)
add_compile_options(-fbuiltin-module-map)
add_compile_options(-fimplicit-module-maps)
add_compile_options(-fprebuilt-module-path=${PREBUILT_MODULE_PATH})
add_module(helloworld helloworld.cpp)
add_executable(main
main.cpp
helloworld.cpp
)
add_dependencies(main helloworld.pcm)
Assuming that you're using gcc 11 with a Makefile generator, the following code should work even without CMake support for C++20:
cmake_minimum_required(VERSION 3.19) # Lower versions should also be supported
project(cpp20-modules)
# Add target to build iostream module
add_custom_target(std_modules ALL
COMMAND ${CMAKE_COMMAND} -E echo "Building standard library modules"
COMMAND g++ -fmodules-ts -std=c++20 -c -x c++-system-header iostream
WORKING_DIRECTORY ${CMAKE_BINARY_DIR}
)
# Function to set up modules in GCC
function (prepare_for_module TGT)
target_compile_options(${TGT} PUBLIC -fmodules-ts)
set_property(TARGET ${TGT} PROPERTY CXX_STANDARD 20)
set_property(TARGET ${TGT} PROPERTY CXX_EXTENSIONS OFF)
add_dependencies(${TGT} std_modules)
endfunction()
# Program name and sources
set (TARGET prog)
set (SOURCES main.cpp)
set (MODULES mymod.cpp)
# Setup program modules object library
set (MODULE_TARGET prog-modules)
add_library(${MODULE_TARGET} OBJECT ${MODULES})
prepare_for_module(${MODULE_TARGET})
# Setup executable
add_executable(${TARGET} ${SOURCES})
prepare_for_module(${TARGET})
# Add modules to application using object library
target_link_libraries(${TARGET} PRIVATE ${MODULE_TARGET})
Some explanation:
A custom target is added to build the standard library modules, in case you want to include standard library header units (search for "Standard Library Header Units" here). For simplicity, I just added iostream here.
Next, a function is added to conveniently enable C++20 and Modules TS for targets
We first create an object library to build the user modules
Finally, we create our executable and link it to the object library created in the previous step.
Not consider the following main.cpp:
import mymod;
int main() {
helloModule();
}
and mymod.cpp:
module;
export module mymod;
import <iostream>;
export void helloModule() {
std::cout << "Hello module!\n";
}
Using the above CMakeLists.txt, your example should compile fine (successfully tested in Ubuntu WSL with gcc 1.11.0).
Update:
Sometimes when changing the CMakeLists.txt and recompiling, you may encounter an error
error: import "/usr/include/c++/11/iostream" has CRC mismatch
Probably the reason is that every new module will attempt to build the standard library modules, but I'm not sure. Unfortunately I didn't find a proper solution to this (avoiding rebuild if the gcm.cache directory already exists is bad if you want to add new standard modules, and doing it per-module is a maintenance nightmare). My Q&D solution is to delete ${CMAKE_BINARY_DIR}/gcm.cache and rebuild the modules. I'm happy for better suggestions though.
CMake ships with experimental support for C++20 modules:
https://gitlab.kitware.com/cmake/cmake/-/blob/master/Help/dev/experimental.rst
This is tracked in this issue:
https://gitlab.kitware.com/cmake/cmake/-/issues/18355
There is also a CMakeCXXModules repository that adds support for modules to CMake.
https://github.com/NTSFka/CMakeCxxModules
While waiting for proper C++20 modules support in CMake, I've found that if using MSVC Windows, for right now you can make-believe it's there by hacking around the build instead of around CMakeLists.txt: continously generate with latest VS generator, and open/build the .sln with VS2020. The IFC dependency chain gets taken care of automatically (import <iostream>; just works). Haven't tried Windows clang or cross-compiling. It's not ideal but for now at least another decently workable alternative today, so far.
Important afterthought: use .cppm and .ixx extensions.
CMake does not currently support C++20 modules like the others have stated. However, module support for Fortran is very similar, and perhaps this could be easily changed to support modules in C++20.
http://fortranwiki.org/fortran/show/Build+tools
Now, perhaps there i an easy way to modify this to support C++20 directly. Not sure. It is worth exploring and doing a pull request should you resolve it.
Add MSVC version (revised from #warchantua 's answer):
cmake_minimum_required(VERSION 3.16)
project(Cpp20)
set(CMAKE_CXX_STANDARD 20)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_EXTENSIONS OFF)
set(PREBUILT_MODULE_DIR ${CMAKE_BINARY_DIR}/modules)
set(STD_MODULES_DIR "D:/MSVC/VC/Tools/MSVC/14.29.30133/ifc/x64") # macro "$(VC_IFCPath)" in MSVC
function(add_module name)
file(MAKE_DIRECTORY ${PREBUILT_MODULE_DIR})
add_custom_target(${name}.ifc
COMMAND
${CMAKE_CXX_COMPILER}
/std:c++latest
/stdIfcDir ${STD_MODULES_DIR}
/experimental:module
/c
/EHsc
/MD
${CMAKE_CURRENT_SOURCE_DIR}/${ARGN}
/module:export
/ifcOutput
${PREBUILT_MODULE_DIR}/${name}.ifc
/Fo${PREBUILT_MODULE_DIR}/${name}.obj
)
endfunction()
set(CUSTOM_MODULES_DIR ${CMAKE_CURRENT_SOURCE_DIR}/modules)
add_module(my_module ${CUSTOM_MODULES_DIR}/my_module.ixx)
add_executable(test
test.cpp
)
target_compile_options(test
BEFORE
PRIVATE
/std:c++latest
/experimental:module
/stdIfcDir ${STD_MODULES_DIR}
/ifcSearchDir ${PREBUILT_MODULE_DIR}
/reference my_module=${PREBUILT_MODULE_DIR}/my_module.ifc
/EHsc
/MD
)
target_link_libraries(test ${PREBUILT_MODULE_DIR}/my_module.obj)
add_dependencies(test my_module.ifc)
With C++20 Modules the file compilation order matters, which is totally new. That's why the implementation is complicated and still experimental in 2023. Please read the authors blogpost
I was not able to find Cmake support for modules. Here is an example how to use modules using clang. I am using Mac and this example works ok on my system. It took me quite a while to figure this out so unsure how general this is across linux or Windows.
Source code in file driver.cxx
import hello;
int main() { say_hello("Modules"); }
Source code in file hello.cxx
#include <iostream>
module hello;
void say_hello(const char *n) {
std::cout << "Hello, " << n << "!" << std::endl;
}
Source code in file hello.mxx
export module hello;
export void say_hello (const char* name);
And to compile the code with above source files, here are command lines on terminal
clang++ \
-std=c++2a \
-fmodules-ts \
--precompile \
-x c++-module \
-Xclang -fmodules-embed-all-files \
-Xclang -fmodules-codegen \
-Xclang -fmodules-debuginfo \
-o hello.pcm hello.mxx
clang++ -std=c++2a -fmodules-ts -o hello.pcm.o -c hello.pcm
clang++ -std=c++2a -fmodules-ts -x c++ -o hello.o \
-fmodule-file=hello.pcm -c hello.cxx
clang++ -std=c++2a -fmodules-ts -x c++ -o driver.o \
-fmodule-file=hello=hello.pcm -c driver.cxx
clang++ -o hello hello.pcm.o driver.o hello.o
and to get clean start on next compile
rm -f *.o
rm -f hello
rm -f hello.pcm
expected output
./hello
Hello, Modules!
Hope this helps, all the best.
Related
I have a c++ CMAKE (VERSION 3.10.2 -std=c++17 ) project that i am able to compile and link with bought gcc and clang. Bought of them produce the target binaries which work as expected. Recently i decided to to try and add another target i.e. webassembly. The project is compiling as expect, however when the EMscripten build is being executed i.e. in the linking phase i get the following error:
Elapsed time: 1 s. (time), 0.002241 s. (clock)
[100%] Linking CXX executable wasmExec.js
cd /Projects/time/time.cpp/build/src/wasm && /usr/bin/cmake -E cmake_link_script CMakeFiles/wasmExec.dir/link.txt --verbose=1
/Projects/emscripten/emsdk/emscripten/1.38.12/em++ -s WASM=1 -s NO_EXIT_RUNTIME=1 -s VERBOSE=1 --pre-js /Projects/time/time.cpp/src/wasm/preModule.js -s DEMANGLE_SUPPORT=1 -s DISABLE_EXCEPTION_CATCHING=0 -s ERROR_ON_UNDEFINED_SYMBOLS=0 #CMakeFiles/wasmExec.dir/objects1.rsp -o wasmExec.js #CMakeFiles/wasmExec.dir/linklibs.rsp
error: Linking globals named '_ZTVN9timeproto3time8defaults20TimeDefaultParametersE': symbol multiply defined!
WARNING:root:Note: Input file "#/tmp/tmpUeJ6zc.response" did not exist.
ERROR:root:Failed to run llvm optimizations:
When i do
c++filt _ZTVN9timeproto3time8defaults20TimeDefaultParametersE
i get
vtable for timeproto::time::defaults::TimeDefaultParameters
from another answer by Stackoverflow i.e.
Possible reasons for symbol multiply defined other than 'extern'
i do understand that i have defined this class more then once, however my problem is that i can not locate that place where i have made the mistake with the second definition. In the previous answer the person had the hint i.e. the cpp file where that he has made that mistake but in my case emscipten is not so generous.
This class is used all over the code base in many many files and after long manual searching i was not able to find anything that can point me at least to the localtion of the second definition. Thus i was hoping that someone can help me with the following questions
1) how can this be troubleshoot further in order to find where exactly the second defintion of the class is occuring, maybe a flag by gcc or clang ?
2) why this error is beeing displayed only when I am trying to compile/build the webassmbly target. The regular Linux64 build target is successefull and the test are also working correctly.
3) I am running cmake with following "add_definitions" i.e.
if(UNIX)
add_definitions(" -pedantic -pedantic-errors -W ")
add_definitions(" -Wall -Wextra -Werror -Wshadow -Wnon-virtual-dtor ")
add_definitions(" -v ")
# add_definitions(" -Worl-style-cast -Wcast-align ")
# add_definitions(" -Wunused -Woverloaded-virtual ")
add_definitions(" -g ")
endif(UNIX)
if the TimeDefaultParameters has been defined more then once should't clang not complain also for linux build with the above "add_definitions" ?
here is the code below TimeDefaultParameters.cpp This is a very simple file that does not contain any object instead it has 43 "static const uint32_t" variables.
#include "TimeDefaultParameters.h"
namespace timeproto::time::defaults
{
TimeDefaultParameters::TimeDefaultParameters() {
}
TimeDefaultParameters::~TimeDefaultParameters() {
}
const uint32_t TimeDefaultParameters::SIGNED_SHORT_MAX_VALUE = 32767;
.... (another 42 static const uint32_t)
}
and the header file TimeDefaultParameters.h:
#ifndef _TIME_DEFAULT_PARAMETERS_
#define _TIME_DEFAULT_PARAMETERS_
#include <stdint.h>
namespace timeproto::time::defaults
{
class TimeDefaultParameters final
{
public:
explicit TimeDefaultParameters();
virtual ~TimeDefaultParameters();
static const uint32_t SIGNED_SHORT_MAX_VALUE;
.....
.... (another 42 static const uint32_t)
};
}
#endif //#ifndef _TIME_DEFAULT_PARAMETERS_
in cmake i have the set my target properties like:
set_target_properties(wasmExec PROPERTIES LINK_FLAGS "-s WASM=1 -s NO_EXIT_RUNTIME=1 -s VERBOSE=1 --pre-js /Projects/time/time.cpp/src/wasm/preModule.js -s DEMANGLE_SUPPORT=1 -s DISABLE_EXCEPTION_CATCHING=0 -s ERROR_ON_UNDEFINED_SYMBOLS=0" )
this is how i am calling cmake to make the build from withing the build directory
emconfigure cmake -DCMAKE_BUILD_TYPE=Emscripten -G "Unix Makefiles" -DCMAKE_TOOLCHAIN_FILE=/Projects/emscripten/emsdk/emscripten/1.38.12/cmake/Modules/Platform/Emscripten.cmake ../
make -j8
any ideas are greatly appreciated.
ADDITION: 05 January 2020
I was able to find a workaround for this problem but i still it leaves some questions of the nature of the error.
The class in question was part of the archive that was created and loaded dynamically i.e. i had used in the CMAKE part for this library "set(LIB_TYPE SHARED)".
here is full example how cmake generated that archive i.e. the CMakeLists.txt.
set( TIME_DEFAULTS_SRC
...
TimeDefaultParameters.h TimeDefaultParameters.cpp
...
)
set(LIB_TYPE STATIC)
#set(LIB_TYPE SHARED)
add_library(time_defaults ${LIB_TYPE} ${TIME_DEFAULTS_SRC} )
target_include_directories(time_defaults PUBLIC "${CMAKE_CURRENT_SOURCE_DIR}/")
I have changed from dynamic to static and i was able to create the wasm no errors were shown. During the compilation i also saw somewhere in between of the compilation process some warrning i.e. :
WARNING:root:When Emscripten compiles to a typical native suffix for shared libraries (.so, .dylib, .dll) then it emits an LLVM bitcode file. You should then compile that to an emscripten SIDE_MODULE (using that flag) with suffix .wasm (for wasm) or .js (for asm.js). (You may also want to adapt your build system to emit the more standard suffix for a file with LLVM bitcode, '.bc', which would avoid this warning.)
this warrning is now gone. But it is very easy to oversee thing like that especially if the compilation process is taking long time. However my understanding is that the very first error message tells us , "look you have made duplicate definition of some symbol in your code go find the place and make sure that the class is define only once". That was exactly what i was doing i.e. searching the code base for that duplicate definition. Thus now the question is: Why emscripten have a problem does with dynamic linking i.e. i know that it is officially supported i.e.
https://webassembly.org/docs/dynamic-linking/
and is that the source of the error at all or is something else?
Why this error disappears when i change to static. I can reproduce this by simply changing library type!
I think i already found the answer here
https://github.com/emscripten-core/emscripten/wiki/Linking
So the solution in my case was to find the occurrences in the CMAKE file where the library was added dynamically and change that to static linking i.e.
#set(LIB_TYPE SHARED)
set(LIB_TYPE STATIC)
Clang and MSVC already supports Modules TS from unfinished C++20 standard.
Can I build my modules based project with CMake or other build system and how?
I tried build2, it supports modules and it works very well, but i have a question about it's dependency management (UPD: question is closed).
CMake currently does not support C++20 modules.
See also the relevant issue in the CMake issue tracker. Note that supporting modules requires far more support from the build system than inserting a new compiler option. It fundamentally changes how dependencies between source files have to be handled during the build: In a pre-modules world all cpp source files can be built independently in any order. With modules that is no longer true, which has implications not only for CMake itself, but also for the downstream build system.
Take a look at the CMake Fortran modules paper for the gory details. From a build system's point of view, Fortran's modules behave very similar to the C++20 modules.
Update: CMake 3.20 introduces experimental support for Modules with the Ninja Generator (and only for Ninja). Details can be found in the respective pull request. At this stage, this feature is still highly experimental and not intended for production use. If you intend to play around with this anyway, you really should be reading both the Fortran modules paper and the dependency format paper to understand what you're getting into.
This works on Linux Manjaro (same as Arch), but should work on any Unix OS. Of course, you need to build with new clang (tested with clang-10).
helloworld.cpp:
export module helloworld;
import <cstdio>;
export void hello() { puts("Hello world!"); }
main.cpp:
import helloworld; // import declaration
int main() {
hello();
}
CMakeLists.txt:
cmake_minimum_required(VERSION 3.16)
project(main)
set(CMAKE_CXX_STANDARD 20)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_EXTENSIONS OFF)
set(PREBUILT_MODULE_PATH ${CMAKE_BINARY_DIR}/modules)
function(add_module name)
file(MAKE_DIRECTORY ${PREBUILT_MODULE_PATH})
add_custom_target(${name}.pcm
COMMAND
${CMAKE_CXX_COMPILER}
-std=c++20
-stdlib=libc++
-fmodules
-c
${CMAKE_CURRENT_SOURCE_DIR}/${ARGN}
-Xclang -emit-module-interface
-o ${PREBUILT_MODULE_PATH}/${name}.pcm
)
endfunction()
add_compile_options(-fmodules)
add_compile_options(-stdlib=libc++)
add_compile_options(-fbuiltin-module-map)
add_compile_options(-fimplicit-module-maps)
add_compile_options(-fprebuilt-module-path=${PREBUILT_MODULE_PATH})
add_module(helloworld helloworld.cpp)
add_executable(main
main.cpp
helloworld.cpp
)
add_dependencies(main helloworld.pcm)
Assuming that you're using gcc 11 with a Makefile generator, the following code should work even without CMake support for C++20:
cmake_minimum_required(VERSION 3.19) # Lower versions should also be supported
project(cpp20-modules)
# Add target to build iostream module
add_custom_target(std_modules ALL
COMMAND ${CMAKE_COMMAND} -E echo "Building standard library modules"
COMMAND g++ -fmodules-ts -std=c++20 -c -x c++-system-header iostream
WORKING_DIRECTORY ${CMAKE_BINARY_DIR}
)
# Function to set up modules in GCC
function (prepare_for_module TGT)
target_compile_options(${TGT} PUBLIC -fmodules-ts)
set_property(TARGET ${TGT} PROPERTY CXX_STANDARD 20)
set_property(TARGET ${TGT} PROPERTY CXX_EXTENSIONS OFF)
add_dependencies(${TGT} std_modules)
endfunction()
# Program name and sources
set (TARGET prog)
set (SOURCES main.cpp)
set (MODULES mymod.cpp)
# Setup program modules object library
set (MODULE_TARGET prog-modules)
add_library(${MODULE_TARGET} OBJECT ${MODULES})
prepare_for_module(${MODULE_TARGET})
# Setup executable
add_executable(${TARGET} ${SOURCES})
prepare_for_module(${TARGET})
# Add modules to application using object library
target_link_libraries(${TARGET} PRIVATE ${MODULE_TARGET})
Some explanation:
A custom target is added to build the standard library modules, in case you want to include standard library header units (search for "Standard Library Header Units" here). For simplicity, I just added iostream here.
Next, a function is added to conveniently enable C++20 and Modules TS for targets
We first create an object library to build the user modules
Finally, we create our executable and link it to the object library created in the previous step.
Not consider the following main.cpp:
import mymod;
int main() {
helloModule();
}
and mymod.cpp:
module;
export module mymod;
import <iostream>;
export void helloModule() {
std::cout << "Hello module!\n";
}
Using the above CMakeLists.txt, your example should compile fine (successfully tested in Ubuntu WSL with gcc 1.11.0).
Update:
Sometimes when changing the CMakeLists.txt and recompiling, you may encounter an error
error: import "/usr/include/c++/11/iostream" has CRC mismatch
Probably the reason is that every new module will attempt to build the standard library modules, but I'm not sure. Unfortunately I didn't find a proper solution to this (avoiding rebuild if the gcm.cache directory already exists is bad if you want to add new standard modules, and doing it per-module is a maintenance nightmare). My Q&D solution is to delete ${CMAKE_BINARY_DIR}/gcm.cache and rebuild the modules. I'm happy for better suggestions though.
CMake ships with experimental support for C++20 modules:
https://gitlab.kitware.com/cmake/cmake/-/blob/master/Help/dev/experimental.rst
This is tracked in this issue:
https://gitlab.kitware.com/cmake/cmake/-/issues/18355
There is also a CMakeCXXModules repository that adds support for modules to CMake.
https://github.com/NTSFka/CMakeCxxModules
While waiting for proper C++20 modules support in CMake, I've found that if using MSVC Windows, for right now you can make-believe it's there by hacking around the build instead of around CMakeLists.txt: continously generate with latest VS generator, and open/build the .sln with VS2020. The IFC dependency chain gets taken care of automatically (import <iostream>; just works). Haven't tried Windows clang or cross-compiling. It's not ideal but for now at least another decently workable alternative today, so far.
Important afterthought: use .cppm and .ixx extensions.
CMake does not currently support C++20 modules like the others have stated. However, module support for Fortran is very similar, and perhaps this could be easily changed to support modules in C++20.
http://fortranwiki.org/fortran/show/Build+tools
Now, perhaps there i an easy way to modify this to support C++20 directly. Not sure. It is worth exploring and doing a pull request should you resolve it.
Add MSVC version (revised from #warchantua 's answer):
cmake_minimum_required(VERSION 3.16)
project(Cpp20)
set(CMAKE_CXX_STANDARD 20)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_EXTENSIONS OFF)
set(PREBUILT_MODULE_DIR ${CMAKE_BINARY_DIR}/modules)
set(STD_MODULES_DIR "D:/MSVC/VC/Tools/MSVC/14.29.30133/ifc/x64") # macro "$(VC_IFCPath)" in MSVC
function(add_module name)
file(MAKE_DIRECTORY ${PREBUILT_MODULE_DIR})
add_custom_target(${name}.ifc
COMMAND
${CMAKE_CXX_COMPILER}
/std:c++latest
/stdIfcDir ${STD_MODULES_DIR}
/experimental:module
/c
/EHsc
/MD
${CMAKE_CURRENT_SOURCE_DIR}/${ARGN}
/module:export
/ifcOutput
${PREBUILT_MODULE_DIR}/${name}.ifc
/Fo${PREBUILT_MODULE_DIR}/${name}.obj
)
endfunction()
set(CUSTOM_MODULES_DIR ${CMAKE_CURRENT_SOURCE_DIR}/modules)
add_module(my_module ${CUSTOM_MODULES_DIR}/my_module.ixx)
add_executable(test
test.cpp
)
target_compile_options(test
BEFORE
PRIVATE
/std:c++latest
/experimental:module
/stdIfcDir ${STD_MODULES_DIR}
/ifcSearchDir ${PREBUILT_MODULE_DIR}
/reference my_module=${PREBUILT_MODULE_DIR}/my_module.ifc
/EHsc
/MD
)
target_link_libraries(test ${PREBUILT_MODULE_DIR}/my_module.obj)
add_dependencies(test my_module.ifc)
With C++20 Modules the file compilation order matters, which is totally new. That's why the implementation is complicated and still experimental in 2023. Please read the authors blogpost
I was not able to find Cmake support for modules. Here is an example how to use modules using clang. I am using Mac and this example works ok on my system. It took me quite a while to figure this out so unsure how general this is across linux or Windows.
Source code in file driver.cxx
import hello;
int main() { say_hello("Modules"); }
Source code in file hello.cxx
#include <iostream>
module hello;
void say_hello(const char *n) {
std::cout << "Hello, " << n << "!" << std::endl;
}
Source code in file hello.mxx
export module hello;
export void say_hello (const char* name);
And to compile the code with above source files, here are command lines on terminal
clang++ \
-std=c++2a \
-fmodules-ts \
--precompile \
-x c++-module \
-Xclang -fmodules-embed-all-files \
-Xclang -fmodules-codegen \
-Xclang -fmodules-debuginfo \
-o hello.pcm hello.mxx
clang++ -std=c++2a -fmodules-ts -o hello.pcm.o -c hello.pcm
clang++ -std=c++2a -fmodules-ts -x c++ -o hello.o \
-fmodule-file=hello.pcm -c hello.cxx
clang++ -std=c++2a -fmodules-ts -x c++ -o driver.o \
-fmodule-file=hello=hello.pcm -c driver.cxx
clang++ -o hello hello.pcm.o driver.o hello.o
and to get clean start on next compile
rm -f *.o
rm -f hello
rm -f hello.pcm
expected output
./hello
Hello, Modules!
Hope this helps, all the best.
I want to integrate clang-tidy to our C and C++, CMake based project which is compiled using a custom GCC toolchain.
I've tried following this tutorial, setting CMAKE_CXX_CLANG_TIDY. I've also tried generating a compilation database by setting CMAKE_EXPORT_COMPILE_COMMANDS to ON and pointing run-clang-tidy.py to its directory.
In both cases, I've encountered (the same) few errors that are probably related to differences between Clang and GCC:
Some warning flags that are enabled in the CMake files are not supported in Clang but are supported in GCC (like -Wlogical-op). As the compiler is GCC, the file builds correctly, and the flag is written to the compilation database, but clang-tidy complains about it.
clang-tidy complains some defines and functions are unavailable, even though the code compiles just fine. As an example, the android-cloexec-open check suggested using O_CLOEXEC to improve security and force the closing of files, but trying to use this define leads to an undefined identifier error (even though our GCC compiles the code).
As an example to a function that is not found, there is clock_gettime.
Our code compiles with the C11 standard and C++14 standard, without GNU extensions:
set(CMAKE_C_STANDARD 11)
set(CMAKE_CXX_STANDARD 14)
set(CMAKE_C_EXTENSIONS OFF)
set(CMAKE_CXX_EXTENSIONS OFF)
The custom toolchain is a cross-compilation toolchain which runs on Linux and compiles to FreeBSD.
Is there a way to disable the passing of some flags by CMake to clang-tidy? Am I using clang-tidy wrong?
I suspect this issue is related to disabling GNU extensions, using a cross-compilation toolchain, and some feature-test-macro which is not defined by default in Clang but is defined with GCC (e.g. _GNU_SOURCE/_POSIX_SOURCE). If this is the case, how can I check it? If not, should I use clang-tidy differently?
EDIT
As #pablo285 asked, here are 2 warnings I get for a single file, and then as I added --warnings-as-errors=*, the build stops:
error: unknown warning option '-Wlogical-op' ; did you mean '-Wlong-long'? [clang-diagnostic-error]
<file path>: error: use of undeclared identifier 'O_CLOEXEC' [clang-diagnostic-error]
O_WRONLY | O_CLOEXEC
^
I decided to write a python script that will replace clang-tidy, receive the commandline from CMake and edit it to fix various errors. Here are the modification to the commandline I tried:
Remove none clang compile flags
This helps with things like the first warning, as now I don't pass flags that clang doesn't know. It seems like I can't configure CMake to pass different set of flags to GCC and to clang-tidy, so if anyone is familiar with some solution to this problem, I'll be happy to hear!
I changed the include directories that are passed to clang-tidy
As mentioned in the post, I use a custom toolchain (which cross-compiles). I used this post and Python to extract the list of standard include directories, and added them to the flag list as a list of -isystem <dir>. I also added -nostdinc so that clang-tidy won't try to look on his own headers instead of mine
This helped with the issue above, as now various defines such as O_CLOEXEC is defined in the toolchain's headers, but as my toolchain is based on GCC, clang couldn't parse the <type_traits> header which includes calls to many compiler intrinsics
I'm not sure what's the best approach in this case
#shycha: Thanks for the tip, I'll try disabling this specific check and I'll edit this post again
Ok, I think that I have a solution. After a couple of evenings I was able to make it work.
In general I compile like this
rm -rf build
mkdir build
cd build
cmake -C ../cmake-scripts/clang-tidy-all.cmake .. && make
Where cmake-scripts directory contains:
clang-tidy-all.cmake
toolchain_arm_clang.cmake
The two important files are listed below.
But what is more important, is how you need to compile this.
First, toolchain_arm_clang.cmake is referenced directly from clang-tidy-all.cmake via set(CMAKE_TOOLCHAIN_FILE ...). It must be, however, referenced from the point of view of the building directory, so if you use multiple levels of build-dirs, e.g.: build/x86, build/arm, build/darwin, etc., then you must modify that path accordingly.
Second, the purpose of set(CONFIG_SCRIPT_PRELOADED ...) is to be sure that the config script was pre-loaded, i.e., cmake -C ../cmake-scripts/clang-tidy-all.cmake ..
Typically, you would want to have something like this somewhere in your CMakeLists.txt file:
message(STATUS "CONFIG_SCRIPT_PRELOADED: ${CONFIG_SCRIPT_PRELOADED}")
if(NOT CONFIG_SCRIPT_PRELOADED)
message(FATAL_ERROR "Run cmake -C /path/to/cmake.script to preload a config script!")
endif()
Third, there is /lib/ld-musl-armhf.so.1 hard-coded in set(CMAKE_LINKER_ARM_COMPAT_STATIC ...); on the development box that I use, it points to /lib/libc.so, so it might by OK to use /lib/libc.sh instead. I've never tried.
Fourth, using set(CMAKE_C_LINK_EXECUTABLE ...) and set(CMAKE_LINKER_ARM_COMPAT_STATIC ...) was because CMake was complaining about some linking problems during checking the compiler, i.e., before even running make.
Fifth, I was only compiling C++ code, so if you need to compile some C, then it might be required to also properly configure set(CMAKE_C_CREATE_SHARED_LIBRARY ...), but I have no idea whether there is such a config option.
General Advice
Do not integrate it immediately. First test some simple CMake project with one library (preferably a C++ one) and make it work, then add the second library, but in C, tweak it again. And only after that incorporate it into the code base.
Toolchain
I used a custom toolchain with GCC 8.3.0 and musl C library, so locations of some files might be different for other toolchains.
Custom CMake
Some variables, like (already mentioned) CONFIG_SCRIPT_PRELOADED, EXPORT_PACKAGE_TO_GLOBAL_REGISTRY, DO_NOT_BUILD_TESTS, or DO_NOT_BUILD_BENCHMARKS are not generic CMake options, i.e., I use them only in my CMakeLists.txt, so you can safely ignore them.
Variables that are unset at the end of each *.cmake file, e.g., build_test, extra_clang_tidy_unchecks_for_tests_only, don't need to be present in the project's main CMakeLists.txt.
Clang
$ clang --version
clang version 10.0.0 (https://github.com/llvm/llvm-project.git 4650b2f36949407ef25686440e3d65ac47709deb)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /opt/local/bin
Files
clang-tidy-all.cmake:
set(ALL_CXX_WARNING_FLAGS --all-warnings -Weverything -Wno-c++98-compat -Wno-c++98-c++11-compat -Wno-c++98-c++11-c++14-compat -Wno-padded -Wno-c++98-compat-pedantic)
set(CXX_COMPILE_OPTIONS "-std=c++17;-O3;${ALL_CXX_WARNING_FLAGS}" CACHE INTERNAL "description")
set(CMAKE_CROSSCOMPILING True)
set(CMAKE_TOOLCHAIN_FILE "../cmake-scripts/toolchain_arm_clang.cmake" CACHE FILEPATH "CMake toolchain file")
set(CONFIG_SCRIPT_PRELOADED true CACHE BOOL "Ensures that config script was preloaded")
set(build_test False)
if(build_test)
message(STATUS "Using test mode clang-tidy checks!")
set(extra_clang_tidy_unchecks_for_tests_only ",-google-readability-avoid-underscore-in-googletest-name,-cppcoreguidelines-avoid-magic-numbers,-cppcoreguidelines-special-member-functions")
endif()
set(CMAKE_CXX_CLANG_TIDY "clang-tidy;--enable-check-profile;--checks=-*,abseil-string-find-startswith,bugprone-*,cert-*,clang-analyzer-*,cppcoreguidelines-*,google-*,hicpp-*,llvm-*,misc-*,modernize-*,-modernize-use-trailing-return-type,performance-*,readability-*,-readability-static-definition-in-anonymous-namespace,-readability-simplify-boolean-expr,portability-*${extra_clang_tidy_unchecks_for_tests_only}" CACHE INTERNAL "clang-tidy")
message(STATUS "build_test: ${build_test}")
message(STATUS "extra_clang_tidy_unchecks_for_tests_only: ${extra_clang_tidy_unchecks_for_tests_only}")
message(STATUS "CMAKE_CXX_CLANG_TIDY: ${CMAKE_CXX_CLANG_TIDY}")
# We want to skip building tests when clang-tidy is run (it takes too much time and serves nothing)
if(DEFINED CMAKE_CXX_CLANG_TIDY AND NOT build_test)
set(DO_NOT_BUILD_TESTS true CACHE BOOL "Turns OFF building tests")
set(DO_NOT_BUILD_BENCHMARKS true CACHE BOOL "Turns OFF building benchmarks")
endif()
unset(build_test)
unset(extra_clang_tidy_unchecks_for_tests_only)
set(EXPORT_PACKAGE_TO_GLOBAL_REGISTRY "OFF" CACHE INTERNAL "We don't export clang-tidy-all version to global register")
toolchain_arm_clang.cmake:
set(CMAKE_SYSTEM_NAME Linux)
set(CMAKE_SYSTEM_VERSION 4.14.0)
set(CMAKE_SYSTEM_PROCESSOR arm)
set(gcc_version 8.3.0)
set(x_tools "/opt/zynq/xtl")
set(CMAKE_C_COMPILER "clang" CACHE INTERNAL STRING)
set(CMAKE_CXX_COMPILER "clang++" CACHE INTERNAL STRING)
set(CMAKE_RANLIB "llvm-ranlib" CACHE INTERNAL STRING)
set(CMAKE_AR "llvm-ar" CACHE INTERNAL STRING)
set(CMAKE_AS "llvm-as" CACHE INTERNAL STRING)
set(CMAKE_LINKER "ld.lld" CACHE INTERNAL STRING)
execute_process(
COMMAND bash -c "dirname `whereis ${CMAKE_LINKER} | tr -s ' ' '\n' | grep ${CMAKE_LINKER}`"
OUTPUT_VARIABLE cmake_linker_dir
)
string(REGEX REPLACE "\n$" "" cmake_linker_dir "${cmake_linker_dir}")
set(cmake_linker_with_dir "${cmake_linker_dir}/${CMAKE_LINKER}" CACHE INTERNAL STRING)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -iwithsysroot /include/c++/${gcc_version} -iwithsysroot /include/c++/${gcc_version}/arm-linux-musleabihf" CACHE INTERNAL STRING)
set(CMAKE_SYSROOT ${x_tools}/arm-linux-musleabihf)
set(CMAKE_FIND_ROOT_PATH ${x_tools}/arm-linux-musleabihf)
set(CMAKE_INSTALL_PREFIX ${x_tools}/arm-linux-musleabihf)
set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY BOTH)
set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE BOTH)
set(CMAKE_FIND_ROOT_PATH_MODE_PACKAGE NEVER)
set(triple arm-linux-musleabihf)
set(CMAKE_LIBRARY_ARCHITECTURE ${triple})
set(CMAKE_C_COMPILER_TARGET ${triple})
set(CMAKE_CXX_COMPILER_TARGET ${triple})
set(lib_path_arm ${x_tools}/arm-linux-musleabihf/lib)
## Bootstrap library stuff:
set(Scrt1_o ${lib_path_arm}/Scrt1.o)
set(crti_o ${lib_path_arm}/crti.o)
set(crtn_o ${lib_path_arm}/crtn.o)
set(lib_path_gcc ${x_tools}/lib/gcc/${triple}/${gcc_version})
set(crtbeginS_o ${lib_path_gcc}/crtbeginS.o)
set(crtendS_o ${lib_path_gcc}/crtendS.o)
# Clang as linker
# --no-pie disable position independent executable, which is required when building
# statically linked executables.
set(CMAKE_CXX_LINK_EXECUTABLE "clang++ --target=${triple} -Wl,--no-pie --sysroot=${CMAKE_SYSROOT} ${CMAKE_CXX_FLAGS} -fuse-ld=${cmake_linker_with_dir} <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <LINK_LIBRARIES> <OBJECTS> -o <TARGET> ")
set(CMAKE_CXX_CREATE_SHARED_LIBRARY "clang++ -Wl, --target=${triple} --sysroot=${CMAKE_SYSROOT} ${CMAKE_CXX_FLAGS} -fuse-ld=${cmake_linker_with_dir} -shared <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <LINK_LIBRARIES> <OBJECTS> -o <TARGET> ")
#
# Do not use CMAKE_CXX_CREATE_STATIC_LIBRARY -- it is created automatically
# by cmake using ar and ranlib
#
#set(CMAKE_CXX_CREATE_STATIC_LIBRARY "clang++ -Wl,--no-pie,--no-export-dynamic,-v -v --target=${triple} --sysroot=${CMAKE_SYSROOT} ${CMAKE_CXX_FLAGS} -fuse-ld=ld.lld <CMAKE_CXX_LINK_FLAGS> <LINK_FLAGS> <LINK_LIBRARIES> <OBJECTS> -o <TARGET> ")
## Linker as linker
set(CMAKE_LINKER_ARM_COMPAT_STATIC "-pie -EL -z relro -X --hash-style=gnu --eh-frame-hdr -m armelf_linux_eabi -dynamic-linker /lib/ld-musl-armhf.so.1 ${Scrt1_o} ${crti_o} ${crtbeginS_o} -lstdc++ -lm -lgcc_s -lgcc -lc ${crtendS_o} ${crtn_o}")
set(CMAKE_C_LINK_EXECUTABLE "${CMAKE_LINKER} ${CMAKE_LINKER_ARM_COMPAT_STATIC} <CMAKE_C_LINK_FLAGS> <LINK_FLAGS> <LINK_LIBRARIES> <OBJECTS> -o <TARGET>")
# Debian bug 708744(?)
#include_directories("${CMAKE_SYSROOT}/usr/include/")
#include_directories("${CMAKE_SYSROOT}/usr/include/c++/${gcc_version}")
#include_directories("${CMAKE_SYSROOT}/usr/include/c++/${gcc_version}/${triple}")
## Clang workarounds:
set(toolchain_lib_dir_0 "${CMAKE_SYSROOT}/lib")
set(toolchain_lib_dir_1 "${CMAKE_SYSROOT}/../lib")
set(toolchain_lib_dir_2 "${CMAKE_SYSROOT}/../lib/gcc/${triple}/${gcc_version}")
set(CMAKE_TOOLCHAIN_LINK_FLAGS "-L${toolchain_lib_dir_0} -L${toolchain_lib_dir_1} -L${toolchain_lib_dir_2}")
## CMake workarounds
set(CMAKE_EXE_LINKER_FLAGS ${CMAKE_TOOLCHAIN_LINK_FLAGS} CACHE INTERNAL "exe link flags")
set(CMAKE_MODULE_LINKER_FLAGS ${CMAKE_TOOLCHAIN_LINK_FLAGS} CACHE INTERNAL "module link flags")
set(CMAKE_SHARED_LINKER_FLAGS ${CMAKE_TOOLCHAIN_LINK_FLAGS} CACHE INTERNAL "shared link flags")
unset(cmake_linker_with_dir)
unset(cmake_linker_dir)
Maybe not exactly what you're looking for but I'm using this in CMakeLists.txt:
set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
add_custom_target(lint
COMMAND sh -c "run-clang-tidy -header-filter=.* -checks=`tr '\\n' , <${CMAKE_SOURCE_DIR}/checks.txt` >lint.out 2>lint.err"
COMMAND sh -c "grep warning: lint.out || true"
COMMAND ls -lh ${CMAKE_BINARY_DIR}/lint.out
VERBATIM
WORKING_DIRECTORY ${CMAKE_BINARY_DIR}
)
This creates a separate build target (make lint) for the clang-tidy check. clang-tidy takes a long time for my project so I don't want to run it during every build; make lint can be run manually if required, and it's also executed in a CI job after every push to the repo (in a way that makes the CI pipeline fail, blocking the merge, if there are any findings).
The output of make lint is the list of clang-tidy findings with as little context as possible. The full output, including context for findings, is in lint.out, and error messages are in lint.err, both of which I'm saving as CI artefacts.
checks.txt is a text file in the project root that defines which clang-tidy checks to activate, like so:
*
-altera-id-dependent-backward-branch
-altera-struct-pack-align
-altera-unroll-loops
-android-*
The first line enables all available checks, the other lines disable checks that I don't want.
Will only work in a Unix-like system of course.
I am trying to apply Link Time Optimization with LLVM on a CMake Project, that creates a shared library. My question is pretty much the same as this one:
Switching between GCC and Clang/LLVM using CMake.
However, the answers do not seem to be applicable anymore, since llvm-ld is not present in the new versions. On the command line, I run the following commands to get LTO (Assuming there are only 2 .cpp files):
Compile to byte code:
clang++ -c FirstClass.cpp -O3 -flto -o FirstClass.bc
clang++ -c SecondClass.cpp -O3 -flto -o SecondClass.bc
Link byte code:
llvm-link FirstClass.bc SecondClass.bc -o unoptimized.bc
Optimize byte code:
opt -O3 unoptimized.bc -o optimized.bc
Convert byte code to shared object:
clang++ -shared optimized.bc -o libTest.so
Could somebody please tell me how to have CMake run the additional steps?
The correct way to use Clang and enable LTO is using the -flto flag to the clang command line both at compile and link time.
In addition, you will need to be working on a platform with a linker that either directly supports LTO (Apple's platforms generally) or that have an LLVM linker plugin (Linux using the Gold linker, but I think some have gotten the BFD linker to support the linker plugin as well). If you're using the linker plugin, you'll need to make sure your install of LLVM built and installed the plugin. If it did, Clang will automatically add the necessary linker command line options to use the plugin when linking with -flto, even for shared objects.
Also, The LLVM project is working on a new linker (LLD) which will support LTO out of the box on all the platforms it supports, but it is still pretty early days. Currently I know of folks testing out its LTO support on Windows and Linux, and it seems to be working well but still misses many features.
check_ipo_supported() resulted for me in "Policy CMP0069 is not set" error on CMake 3.9.1.
Per its help, CMake up to 3.8 only supported Intel compiler's LTO. It didn't work on XCode 9's clang for me either.
What worked, in the end:
cmake_policy(SET CMP0069 NEW)
include(CheckIPOSupported)
check_ipo_supported()
add_executable(Foobar SOURCES)
set_target_properties(Foobar PROPERTIES INTERPROCEDURAL_OPTIMIZATION TRUE)
Looks like add_executable() needs to be after cmake_policy(SET CMP0069 NEW).
LTO cache
target_link_libraries(Foobar "-Wl,-cache_path_lto,${PROJECT_BINARY_DIR}/lto.cache") did no harm.
Pick your command-line option depending on your linker.
More brutal option
According to #ChandlerCarruth's answer:
if ("${CMAKE_CXX_COMPILER_ID}" MATCHES "Clang")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -flto")
target_link_libraries(Foobar -flto)
endif ()
Enabling (thin) lto on Cmake 3.9 and newer should be straightforward:
include(CheckIPOSupported)
check_ipo_supported()
set_target_properties(myProject PROPERTIES INTERPROCEDURAL_OPTIMIZATION TRUE)
Instead of set_target_properties per project, a single global setting of set(CMAKE_INTERPROCEDURAL_OPTIMIZATION TRUE) can be done.
In order to speed up recompiles, a cache for LTO can be set:
function(append value)
foreach(variable ${ARGN})
set(${variable} "${${variable}} ${value}" PARENT_SCOPE)
endforeach(variable)
endfunction()
append("-fuse-ld=gold -Wl,--no-threads,--plugin-opt,cache-dir=${PROJECT_BINARY_DIR}/lto.cache" CMAKE_EXE_LINKER_FLAGS CMAKE_SHARED_LINKER_FLAGS)
This forces gold as linker, in order to use the right command line options. It might require a symlink of /usr/lib/LLVMgold.so to /usr/lib/llvm-4.0/lib/LLVMgold.so.
I've written the following groundbreaking GPU-powered application:
int main() { return 0; }
and I'm trying to build it using CMake. Here's my CMakeLists.txt file:
cmake_minimum_required(VERSION 2.8)
set(CMAKE_CXX_FLAGS ${CMAKE_CXX_FLAGS} " -std=c++11" )
find_package(CUDA QUIET REQUIRED)
set(CUDA_NVCC_FLAGS ${CUDA_NVCC_FLAGS} " -std=c++11")
cuda_add_executable(a a.cu)
With this file, and for some reason, the compilation fails; it looks like (although I'm not sure) the reason is that CMake is having NVCC use the CUDA include directory twice. The result:
/home/joeuser/opt/cuda/bin/nvcc -M -D__CUDACC__ /home/joeuser/src/kt3/a.cu -o /home/joeuser/src/kt3/CMakeFiles/a.dir//a_generated_a.cu.o.NVCC-depend -ccbin /usr/bin/ccache -m64 --std c++11 -Xcompiler ,\"-g\" -std=c++11 -DNVCC -I/home/joeuser/opt/cuda/include -I/home/joeuser/opt/cuda/include
nvcc fatal : redefinition of argument 'std'
CMake Error at a_generated_a.cu.o.cmake:207 (message):
Error generating
/home/joeuser/src/kt3/CMakeFiles/a.dir//./a_generated_a.cu.o
You might be asking why my MWE is not more terse. Why do I need those option-setting lines above if I don't use C++11 anyway? Well, if I remove them, I still get the double-include, but have an earlier failure involving ccache about which I will probably ask another question.
So is the problem actually the double-include, or is it something else? And what should I do?
Additional information:
I don't have root on this machine.
CMake version: 3.3.2 .
Distribution: Fedora 22 (sorry, I can't help it; but I also seem to be getting this on Debian Stretch as well).
CUDA install location: $HOME/opt/cuda , and its binary directory is in $PATH.
On another system, with a different configuration and distro (Fedora 20, CUDA 7.5 but in another local dir, possibly other differences) I do not get this behavior.
I think the problem is that the nvcc flags are propagated to your c/c++ compiler, so some compiler arguments are effectively passed twice. Try using SET(CUDA_PROPAGATE_HOST_FLAGS OFF)