XCode构建失败“架构x86_64的未定义符号”

Question

已解决：有关OpenCV解决方案的查找，请参考Jvinniec的FAST.H解决方案查找编辑

我查找了几个类似的问题，但尚未找到我的代码的解决方案。 我得到的错误如下所示：

Undefined symbols for architecture x86_64:
"fast9_score(unsigned char const*, int, xy*, int, int)", referenced from:
  detectCornersFast(unsigned char const*, int, int, int, int, int*) in main.o
"fast9_detect(unsigned char const*, int, int, int, int, int*)", referenced from:
  detectCornersFast(unsigned char const*, int, int, int, int, int*) in main.o
"nonmax_suppression(xy const*, int const*, int, int*)", referenced from:
  detectCornersFast(unsigned char const*, int, int, int, int, int*) in main.o
"cv::namedWindow(cv::String const&, int)", referenced from:
  _main in main.o
"cv::GaussianBlur(cv::_InputArray const&, cv::_OutputArray const&, cv::Size_<int>, double, double, int)", referenced from:
  _main in main.o
"cv::Mat::deallocate()", referenced from:
  cv::Mat::release() in main.o
"cv::Mat::copySize(cv::Mat const&)", referenced from:
  cv::Mat::operator=(cv::Mat const&) in main.o
  cv::Mat::Mat(cv::Mat const&) in main.o
"cv::String::deallocate()", referenced from:
  cv::String::~String() in main.o
"cv::String::allocate(unsigned long)", referenced from:
  cv::String::String(char const*) in main.o
  cv::String::String(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) in main.o
"cv::imread(cv::String const&, int)", referenced from:
  _main in main.o
"cv::imshow(cv::String const&, cv::_InputArray const&)", referenced from:
  _main in main.o
"cv::resize(cv::_InputArray const&, cv::_OutputArray const&, cv::Size_<int>, double, double, int)", referenced from:
  _main in main.o
"cv::waitKey(int)", referenced from:
  _main in main.o
"cv::fastFree(void*)", referenced from:
  cv::Mat::~Mat() in main.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)

我在XCode中使用的体系结构是通用的。 我建立并添加了openCV的搜索路径，fast9方法位于fast.h标头中。 我试图删除该项目的派生数据，但没有任何帮助。 这是我的main.cpp代码：

#include <vector>
#include <iostream>
#include <cmath>
#include <string>
#include "opencv2/opencv.hpp"
#include "fast.h"

using namespace std;
using namespace cv;

typedef unsigned char byte;

Mat src; Mat dst; Mat tmp;
int rows; int cols;
byte* image;
xy* nonmax;

byte * matToBytes(Mat image){

int size = image.total() * image.elemSize();
byte * bytes = new byte[size];
std::memcpy(bytes, image.data, size * sizeof(byte));

return bytes;

}

xy* detectCornersFast(const byte* im, int xsize, int ysize, int stride, int b, int* ret_num_corners){

xy* corners;
int num_corners; 
int* scores;    
xy* nonmax;

corners = fast9_detect(im, xsize, ysize, stride, b, &num_corners); 
scores = fast9_score(im, stride, corners, num_corners, b);    
nonmax = nonmax_suppression(corners, scores, num_corners, ret_num_corners);

free(corners);   
free(scores);

return nonmax;

}

int main(int argc, const char * argv[]) {

// insert code here...

vector<Mat> img; 
vector<int> num_corners;
vector<xy*> nms;

src = imread(argv[1],1);

img.push_back(src);
double factor = sqrt(2);

for(int i = 0; i < 5; i++){  

    int ret_num_corners;
    rows = src.rows/factor;
    cols = src.cols/factor;

    resize(src, tmp, Size(rows,cols),0,0, INTER_NEAREST);

    GaussianBlur(tmp, dst, Size(5,5),0,0);

    image = matToBytes(dst);

    nonmax = detectCornersFast(image, tmp.rows, tmp.cols, tmp.rows, 20, &ret_num_corners);

    img.push_back(dst);
    nms.push_back(nonmax);
    num_corners.push_back(ret_num_corners);

    src = dst;

}

for(int i = 0; i <img.size(); i++){

    string name = "Display Window " + std::to_string(i);
    cout << "Number of Corners" << num_corners[i] << endl;
    cout << "Nonmax Supression" << nms[i] << endl;
    namedWindow(name, WINDOW_AUTOSIZE); 
    imshow(name, img[i]);  
}

waitKey();
return 0; 
}

编辑：对于OpenCV，它有助于从本教程中添加链接器标志： OpenCV快速方法来自于我项目文件夹中的.c文件。 .h文件也在那里。 我只是在此发布.h文件，因为.c文件太长了，因为它是机械生成的。

解决方案：快速方法是我的错误。 我完全忘记在main.cpp中的#include "fast.h"处添加Extern "C" {} ，因为它调用C文件而不是C ++文件。

快

#ifndef FAST_H
#define FAST_H

typedef struct { int x, y; } xy; 
typedef unsigned char byte;

int fast9_corner_score(const byte* p, const int pixel[], int bstart);

xy* fast9_detect(const byte* im, int xsize, int ysize, int stride, int b, int* ret_num_corners);

int* fast9_score(const byte* i, int stride, xy* corners, int num_corners, int b);

xy* fast9_detect_nonmax(const byte* im, int xsize, int ysize, int stride, int b, int* ret_num_corners);

xy* nonmax_suppression(const xy* corners, const int* scores, int num_corners, int* ret_num_nonmax);

#endif

Answer 1

听起来好像您只在项目的“ Build Settings”->“ Search Paths”->“ User Header Search Paths”字段下为这些依赖项添加了头搜索路径。 这将使xcode能够知道这些类在您开发时就存在，但不会使您的项目得以构建。 您还应确保您的项目知道在哪里可以找到定义所使用的openCV和fast9符号的库（即， .so文件用于这两个依赖项的位置）。 您可以通过在“构建设置”->“链接”->“其他链接器标志”下添加适当的链接器标志来完成此操作。 还请确保为您要编译的目标（不仅仅是项目）设置了这些字段。

如果您不确定链接器的标志是什么，请参考以下示例。 我想链接以下库：

/path/to/some/project/lib/libfoo.so

我将以下内容添加到“其他链接器标志”字段中：

-L/path/to/some/project/lib -lfoo

基本上， -L在目录之前，而-l在库名称之前，没有'lib'和'.so'。

注意：每当我在xcode项目中包括外部依赖项时，我还将“构建设置”->“搜索路径”->“始终搜索用户路径”设置为“是”，以防万一任何依赖项都包含带有尖括号的标头，例如#include <header.h> 。