File Fixer

file_fixer.cpp

// g++ file_fixer.cpp --std=c++20 -O3 -o file_fixer
// usage: ./file_fixer file_to_fix file_ref_a file_ref_b ....

#include <string_view>
#include <iostream>
#include <span>
#include <vector>
#include <unistd.h>
#include <sys/mman.h>
#include <sys/stat.h>
#include <fcntl.h>

constexpr size_t CHUNK = 1024*1024; //mb

int main_cpp(std::span<const std::string> args) {
  size_t max_size;
  std::vector<std::span<unsigned char>> maps;
  for (const auto arg : args) {
    std::cout << "OPEN " << arg;
    const bool writeable = (arg == args[0]);
    if (writeable) {
       std::cout << " (WRITE)";
    }
    std::cout << std::endl;
    int fd = open(arg.c_str(), writeable ? O_RDWR : O_RDONLY);
    if (fd == -1) {
      std::cerr << "Couln't open file " << arg << std::endl;
      return -1;
    }
    struct stat sb;
    if (fstat(fd, &sb) == -1) {
      std::cerr << "Couln't get file size " << arg << std::endl;
      return -2;
    }
    const size_t size = sb.st_size;
    unsigned char* addr = (unsigned char*)mmap(NULL, size, writeable ? PROT_WRITE : PROT_READ, MAP_SHARED, fd, 0);
    if (addr == MAP_FAILED) {
      std::cerr << "Couln't memory map file " << arg << std::endl;
      return -3;
    }
    max_size = std::max(max_size, size);
    maps.push_back(std::span{
      addr,
      size
    });
  }

  std::cout << "VALIDATE" << std::endl;
  bool fix = false;
  const size_t size = maps[0].size();
  for (size_t i = 0; i < size; ++i) {
    if ((i % CHUNK) == 0) {
      std::cout << "\e[2K" << (i / 1024 / 1024) << " / " << (size / 1024 / 1024) << "MiB" << std::flush << "\r";
    }
    bool first = true;
    unsigned char data;
    unsigned char odata;
    for (const auto map: maps) {
      if (first) {
        first = false;
        data = map[i];
        odata = data;
      }
      if (i >= map.size()) {
        continue;
      }
      if (odata != 0x00) {
        if (map[i] != 0x00 && odata != map[i]) {
          std::cerr << "WRONG DATA" << std::endl;
          return -2;
        }
      } else if (data == 0x00) {
        odata = map[i];
      }
    }
    if (data == 0x00 && odata != 0x00) {
      fix = true;
    }
  }
  std::cout << "\e[2K\r";
  if (!fix) {
    std::cout << "NOTHING TO FIX" << std::endl;
    return 0;
  }

  std::cout << "FIX" << std::endl;
  for (size_t i = 0; i < size; ++i) {
    if ((i % CHUNK) == 0) {
      std::cout << "\e[2K" << (i / 1024 / 1024) << " / " << (size / 1024 / 1024) << "MiB" << std::flush << "\r";
    }
    if (maps[0][i] == 0x00) {
      for (const auto map: maps) {
        if (i >= map.size()) {
          continue;
        }
        if (map[i] != 0x00) {
          maps[0][i] = map[i];
          break;
        }
      }
    }
  }
  std::cout << "\e[2K\r";

  return 0;
}

int main(int argc, const char* argv[]) {
  static std::vector<std::string> args(argv + 1, argv + argc);
  return main_cpp(args);
}