I'm trying to build a python wrapper using the following Makefile:
CC=/usr/local/opt/llvm/bin/clang
OS_NAME=$(shell uname -s)
ifeq ($(OS_NAME),Linux)
LAPACKLDFLAGS=/usr/lib64/atlas/libsatlas.so # single-threaded blas
#LAPACKLDFLAGS=/usr/lib64/atlas/libtatlas.so # multi-threaded blas
#BLAS_THREADING=-D MULTITHREADED_BLAS # remove this if wrong
endif
ifeq ($(OS_NAME),Darwin) # Mac OS X
LAPACKLDFLAGS=-framework Accelerate # for OS X
endif
LAPACKCFLAGS=-Dinteger=int $(BLAS_THREADING)
STATICLAPACKLDFLAGS=-fPIC -Wall -g -fopenmp -static -static-libstdc++ /home/lear/douze/tmp/jpeg-6b/libjpeg.a /usr/lib64/libpng.a /usr/lib64/libz.a /usr/lib64/libblas.a /usr/lib/gcc/x86_64-redhat-linux/4.9.2/libgfortran.a /usr/lib/gcc/x86_64-redhat-linux/4.9.2/libquadmath.a # statically linked version
CFLAGS= -fPIC -Wall -g -std=c++11 $(LAPACKCFLAGS) -fopenmp -DUSE_OPENMP -O3
LDFLAGS=-fPIC -Wall -g -ljpeg -lpng -fopenmp
CPYTHONFLAGS=-I/usr/include/python2.7
SOURCES := $(shell find . -name '*.cpp' ! -name 'deepmatching_matlab.cpp')
OBJ := $(SOURCES:%.cpp=%.o)
HEADERS := $(shell find . -name '*.h')
all: deepmatching
.cpp.o: %.cpp %.h
$(CC) -o $# $(CFLAGS) -c $+
deepmatching: $(HEADERS) $(OBJ)
$(CC) -o $# $^ $(LDFLAGS) $(LAPACKLDFLAGS)
deepmatching-static: $(HEADERS) $(OBJ)
$(CC) -o $# $^ $(STATICLAPACKLDFLAGS)
python: $(HEADERS) $(OBJ)
# swig -python $(CPYTHONFLAGS) deepmatching.i # not necessary, only do if you have swig compiler
/usr/local/opt/llvm/bin/clang $(CFLAGS) -c deepmatching_wrap.c $(CPYTHONFLAGS)
/usr/local/opt/llvm/bin/clang -shared $(LDFLAGS) $(LAPACKLDFLAGS) deepmatching_wrap.o $(OBJ) -o _deepmatching.so $(LIBFLAGS)
clean:
rm -f $(OBJ) deepmatching *~ *.pyc .gdb_history deepmatching_wrap.o _deepmatching.so deepmatching.mex???
Previously, CC was set to g++, however, when I tried to build it like this, I'd get "ERROR: clang: error: unsupported option '-fopenmp".
Now I installed "brew install llvm" as this comes with the -fopenmp option. The unsupported error is resolved for now, but now the compiler doesn't seem to find a header file:
(base) MacBook-Pro-van-Brent:deepmatching BrentDeHauwere$ make python
/usr/local/opt/llvm/bin/clang -o hog.o -fPIC -Wall -g -std=c++11 -Dinteger=int -fopenmp -DUSE_OPENMP -O3 -I/usr/local/opt/llvm/include -c hog.cpp
In file included from hog.cpp:18:
In file included from ./std.h:20:
/usr/local/opt/llvm/bin/../include/c++/v1/math.h:300:15: fatal error: 'math.h' file not found
#include_next <math.h>
^~~~~~~~
1 error generated.
make: *** [hog.o] Error 1
I've tried setting options (I might have set them incorrectly) like -L/usr/local/opt/llvm/lib and -I/usr/local/opt/llvm/include, but no result so far. Any idea how I could point the compiler to the right direction for the header files?
Try running xcode-select —install in your terminal. This installs the xcode command line tools which should also install system headers files (as part of the macos sdk) and set your system include paths.
Related
When I make the Makefile everything works fine, I get a library in the directory dir. And when I run "Make test" I get a testfile that I want to run. But when I want to run this file I get this weird error: ./programma: error while loading shared libraries: libprogramma.so: cannot open shared object file: No such file or directory. I have tried running the program on both WSL and Linux, but nothing makes this error go away. Can anyone help me?
Here I have my Makefile which makes the library and the executable:
INC_DIR = include
SRC_DIR = src
SOURCES = $(sort $(shell find $(SRC_DIR) -name '*.cc'))
OBJECTS = $(SOURCES:.cc=.o)
DEPS = $(OBJECTS:.o=.d)
TARGET = programma
CXX = g++
CFLAGS = -Wall -Wextra -Wpedantic -std=c++11
CPPFLAGS = $(addprefix -I, $(INC_DIR))
.PHONY: all clean debug release
release: CFLAGS += -O3 -DNDEBUG
release: all
debug: CFLAGS += -O0 -DDEBUG -ggdb3
debug: all
all: $(TARGET)
clean:
rm -f $(OBJECTS) $(DEPS) lib/*.so programma *.d
$(TARGET): $(OBJECTS)
$(CXX) $(CFLAGS) $(CPPFLAGS) -fPIC -shared -o lib/lib$#.so $^
-include $(DEPS)
%.o: %.cc
$(CXX) $(CFLAGS) $(CPPFLAGS) -fPIC -MMD -o $# -c $<
test:
$(CXX) $(CFLAGS) -L./lib $(CPPFLAGS) -MMD -o programma tests/main.cc -l$(TARGET)
Executables on Linux don't look for shared libraries in the directory they're located in, at least by default.
You can either fix that at link-time, by passing -Wl,-rpath='$ORIGIN', or at runtime, by setting LD_LIBRARY_PATH env variable to the directory with the library. (LD_LIBRARY_PATH=path/to/lib ./programma)
environment
ubuntu 18.04 (a virtual environment created by vagrant)
gcc5.5.0
I'm trying to build my own little OS.
I'm writing the kernel in C++/NASM.
I'm going to compile each of the several C++ files that make up the kernel and then link them together to create an executable file. (This is because the entry point of my own kernel is not "main", so I dare not link at the same time as compiling)
Error
The compilation of each C++ file will run without any problems and each object file will be created.
However, I get the following error in linking the object files.
But I understand that a ".a" file is a static object, not a dynamic object, so I can not understand why I'm getting this error.
I'd appreciate it if you could let me know.
ld: attempted static link of dynamic object `/usr/lib/x86_64-linux-gnu/libc++abi.a'
Here's the full text of the log
mkdir -p ./obj
g++ -O2 -g -MMD -Wall -I./include -o obj/hoge.o -c src/hoge.cpp
mkdir -p ./obj
g++ -O2 -g -MMD -Wall -I./include -o obj/huga.o -c src/huga.cpp
mkdir -p ./obj
g++ -O2 -g -MMD -Wall -I./include -o obj/piyo.o -c src/piyo.cpp
mkdir -p ./obj
nasm -f elf64 -o obj/assm.o src/assm.asm
ld --entry <ENTRY_POINT> --static -o ./ELF/kernel.elf ./obj/hoge.o ./obj/huga.o ./obj/piyo.o ./obj/assm.o -lc -lc++ -lc++abi -L/usr/lib/x86_64-linux-gnu
ld: attempted static link of dynamic object `/usr/lib/x86_64-linux-gnu/libc++abi.a'
Makefile:12: recipe for target 'ELF/kernel.elf' failed
make: *** [ELF/kernel.elf] Error 1
Makefile
COMPILER = g++
CFLAGS = -O2 -g -MMD -Wall
INCLUDE = -I./include
TARGET = ./ELF/kernel.elf
SRCDIR = ./src
SOURCES := $(shell find $(SRCDIR) -name *.cpp -or -name *.c -or -name *.asm)
OBJDIR = ./obj
OBJECTS = $(OBJDIR)/hoge.o $(OBJDIR)/huga.o $(OBJDIR)/piyo.o $(OBJDIR)/assm.o
DEPENDS = $(OBJECTS:.o=.d)
LDFLAGS = --entry <ENTRY_POINT> --static
$(TARGET): $(OBJECTS)
ld $(LDFLAGS) -o $(TARGET) $(OBJECTS) -lc -lc++ -lc++abi -L/usr/lib/x86_64-linux-gnu
$(OBJDIR)/%.o: $(SRCDIR)/%.cpp
-mkdir -p $(OBJDIR)
$(COMPILER) $(CFLAGS) $(INCLUDE) -o $# -c $<
$(OBJDIR)/%.o: $(SRCDIR)/%.asm
-mkdir -p $(OBJDIR)
nasm -f elf64 -o $# $<
all: clean $(TARGET)
clean:
rm -rf $(OBJECTS) $(DEPENDS) $(TARGET)
-include $(DEPENDS)
I am trying to build a Makefile that will build a shared library with g++ and I find that it is not evaluating the OBJECTS variable. This is on Ubuntu 18.04 and all the files are in the same current directory. Secondly it is completely skipping the source file compilation and proceeding directly to evaluate the linking instruction. As a clarification I am using GNU Make 4.1
Here is what I get when I type make all
g++ -shared -pthread -o tree.so
g++: fatal error: no input files
compilation terminated.
Makefile:12: recipe for target 'tree.so' failed
make: *** [tree.so] Error 1
Here is my Makefile code
CC=g++
CFLAGS = -I/usr/local/include -Wall -std=c++17 -O3 -march=native -Ofast -ftree-vectorize
LIBS=-shared -pthread
SOURCES=$(wildcard *.cpp)
OBJECTS=$(wildcard *.o)
TARGET=tree.so
all:$(TARGET)
$(TARGET) : $(OBJECTS)
$(CC) $(LIBS) -o $(OBJECTS) $(TARGET)
$(OBJECTS):$(SOURCES)
$(CC) -c -g $(CFLAGS) $(SOURCES)
clean:
rm -f $(OBJECTS) $(TARGET)
If you only have the *.cpp files in your directories, then there is not any *.o yet, so your $(wildcard *.o) will expand to nothing.
What you want is to get the *.cpp files and compute the corresponding *.o files:
OBJECTS=$(patsubst %.cpp,%.o,$(SOURCES))
or equivalently:
OBJECTS=$(SOURCES:.cpp=.o)
Now, your compiler command is not the best one, because if you touch any source file all will be compiled. You can use instead:
$(OBJECTS): %.o: %.cpp
$(CC) -c -g $(CFLAGS) $< -o $#
So that only the touched files are actually rebuilt.
Also you have the linking command wrong. It should be:
$(TARGET) : $(OBJECTS)
$(CC) $(LIBS) -o $(TARGET) $(OBJECTS)
because the argument to the -o option is the output file, that is the target.
I have an SDL game I have been working on as my first somewhat real project. I decided to introduce Box2D physics as I was not happy with the collision detection. So I installed it to /usr/local/lib/Box2D and in the folder is Box2D.h and supporting folders. I am using MacVim to code on OSX 10.9.2 to develop, and clang++ compiler from the command line.
In my game code I am just trying to create a simple world to test things out:
#include <Box2D/Box2D.h>
.......
world = new b2World(b2Vec2(0.0,9.81));
My make command finds the library, but errors out trying to build.
$ make clean && make
rm -rf obj bin
clang++ -Wall -c -std=c++11 -I/usr/local/lib src/Ball.cpp -o obj/Ball.o
clang++ -Wall -c -std=c++11 -I/usr/local/lib src/Game.cpp -o obj/Game.o
clang++ -Wall -c -std=c++11 -I/usr/local/lib src/Paddle.cpp -o obj/Paddle.o
clang++ -Wall -c -std=c++11 -I/usr/local/lib src/TextureManager.cpp -o obj/TextureManager.o
clang++ -Wall -c -std=c++11 -I/usr/local/lib src/main.cpp -o obj/main.o
clang++ -framework SDL2 -framework SDL2_image -F /Library/Frameworks -L/usr/local/lib/Box2D obj/Ball.o obj/Game.o obj/Paddle.o obj/TextureManager.o obj/main.o -o bin/game
Undefined symbols for architecture x86_64:
"b2World::b2World(b2Vec2 const&)", referenced from:
Game::init() in Game.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [game] Error 1
And here is my Makefile. Box2D is in /usr/local/lib/Box2D/Box2D.h. I am pretty sure my issue is somewhere in the Makefile.
CXX = clang++
CXXFLAGS = -Wall -c -std=c++11 -I/usr/local/lib
SDL = -framework SDL2 -framework SDL2_image
LDFLAGS = $(SDL) -F /Library/Frameworks -L/usr/local/lib/Box2D
SRC_DIR = src
SOURCES = $(wildcard $(SRC_DIR)/*.cpp)
OBJ_DIR = obj
OBJECTS = $(subst $(SRC_DIR)/, $(OBJ_DIR)/, $(patsubst %.cpp, %.o, $(SOURCES)))
#$(warning $(OBJECTS))
BIN_DIR = bin
EXE = game
# run these no matter what
.PHONY: all clean run
all: $(EXE)
$(EXE): $(OBJECTS)
#mkdir -p $(BIN_DIR)
$(CXX) $(LDFLAGS) $(OBJECTS) -o $(BIN_DIR)/$(EXE)
$(OBJ_DIR)/%.o: $(SRC_DIR)/%.cpp
#mkdir -p $(OBJ_DIR)
$(CXX) $(CXXFLAGS) $< -o $#
clean:
rm -rf obj bin
run:
./$(BIN_DIR)/$(EXE)
It does not appear that you ever actually link the Box2D library anywhere? You use -L to specify directories to search while linking, but I don't see a -l option to actually link the Box2D library (whatever it's called).
Your output line seems to bear this out:
clang++ -framework SDL2 -framework SDL2_image -F /Library/Frameworks -L/usr/local/lib/Box2D obj/Ball.o obj/Game.o obj/Paddle.o obj/TextureManager.o obj/main.o -o bin/game
You need to get -lbox2d (or whatever the correct name for the Box2D library is) in there.
I'm trying to compile a program (which isn't mine):
make -f makefile
... using the following makefile:
# Compiler for .cpp files
CPP = g++
# Use nvcc to compile .cu files
NVCC = nvcc
NVCCFLAGS = -arch sm_20 # For fermi's in keeneland
# Add CUDA Paths
ICUDA = /usr/lib/nvidia-cuda-toolkit/include
LCUDA = /usr/lib/nvidia-cuda-toolkit/lib64
# Add CUDA libraries to the link line
LFLAGS += -lcuda -lcudart -L$(LCUDA) -lgomp
# Include standard optimization flags
CPPFLAGS = -O3 -c -I $(ICUDA) -Xcompiler -fopenmp -DTHRUST_DEVICE_BACKEND=THRUST_DEVICE_BACKEND_OMP
# List of all the objects you need
OBJECTS = timer.o ar1.o kGrid.o vfInit.o parameters.o
# Rule that tells make how to make the program from the objects
main : main.o $(OBJECTS)
$(CPP) -o main main.o $(OBJECTS) $(LFLAGS)
# Rule that tells make how to turn a .cu file into a .o
%.o: %.cu
$(NVCC) ${NVCCFLAGS} $(CPPFLAGS) -c $<
# How does make know how to turn a .cpp into a .o? It's built-in!
# but if you wanted to type it out it would look like:
# %.o: %.cpp
# $(CPP) $(CPPFLAGS) -c $<
clean :
rm -f *.o
rm -f core core.*
veryclean :
rm -f *.o
rm -f core core.*
rm -f main
Which results in the following commands:
nvcc -arch sm_20 -O3 -c -I /usr/lib/nvidia-cuda-toolkit/include -Xcompiler -fopenmp -DTHRUST_DEVICE_BACKEND=THRUST_DEVICE_BACKEND_OMP -c main.cu
g++ -O3 -c -I /usr/lib/nvidia-cuda-toolkit/include -Xcompiler -fopenmp -DTHRUST_DEVICE_BACKEND=THRUST_DEVICE_BACKEND_OMP -c -o timer.o timer.cpp
g++: error: unrecognized command line option â-Xcompilerâ
make: *** [timer.o] Error 1
I don't understand the makefile: the -xCompiler flag (in the variable CPPFLAGS) should be used only by the nvcc compiler, not g++. Therefore, I understand why I am getting an error. However, I don't understand, from my basic understanding of the makefile above, why at some point the variable CPPFLAGS follows g++ (variable CPP). I don't see any such sequence in the makefile.
Your main rule requires timer.o. There is no explicit rule for timer.o so make uses a built in implicit rule (as mentioned in the comment at the end of your makefile). The implicit rule for converting .cpp files into .o files has the form
$(CPP) $(CPPFLAGS) -c $<
So it's compiling using the options in CPPFLAGS which contains -Xcompiler. You probably want the -Xcompiler flag to be in NVCCFLAGS and not CPPFLAGS.