xmlparser/File.h

// Copyright 2017 Patrick Brosi
// info@patrickbrosi.de

#ifndef XML_FILE_H_
#define XML_FILE_H_

#include <cstring>
#include <fstream>
#include <map>
#include <sstream>
#include <stack>
#include <string>

namespace xml {

const static size_t BUFFER_S = 16 * 1024;

class XmlFileException : public std::exception {
 public:
  XmlFileException(std::string msg, std::string file, const char* p, char* buff,
                   size_t offset) {
    std::stringstream ss;
    ss << file << " at position " << (offset + (p - buff)) << ": " << msg;
    _msg = ss.str();
  }
  ~XmlFileException() throw() {}

  virtual const char* what() const throw() { return _msg.c_str(); };

 private:
  std::string _msg;
};

enum State {
  NONE,
  IN_TAG_NAME,
  IN_TAG_NAME_META,
  IN_TAG,
  IN_TAG_CLOSE,
  IN_TAG_NAME_CLOSE,
  IN_TAG_TENTATIVE,
  IN_ATTRKEY,
  AFTER_ATTRKEY,
  AW_IN_ATTRVAL,
  IN_ATTRVAL_SQ,
  IN_ATTRVAL_DQ,
  IN_TEXT,
  IN_COMMENT_TENTATIVE,
  IN_COMMENT_TENTATIVE2,
  IN_COMMENT,
  IN_COMMENT_CL_TENTATIVE,
  IN_COMMENT_CL_TENTATIVE2,
  AW_CLOSING,
  WS_SKIP
};

struct AttrCmp {
  bool operator()(const char* const& a, const char* const& b) const {
    return std::strcmp(a, b) < 0;
  }
};

struct ParserState {
  ParserState() : s(NONE), hanging(0), off(0){};
  std::stack<std::string> tagStack;
  State s;
  size_t hanging;
  int64_t off;
};

typedef std::map<const char*, const char*, AttrCmp> AttrMap;

struct Tag {
  const char* name;
  const char* text;
  AttrMap attrs;
};

class File {
 public:
  File(const std::string& path);
  ~File();

  const Tag& get() const;

  bool next();
  size_t level() const;
  void reset();
  ParserState state();
  void setState(const ParserState& s);
  static std::string decode(const char* str);
  static std::string decode(const std::string& str);

 private:
  int _file;
  ParserState _s;
  ParserState _prevs;
  char** _buffer;
  char* _c;
  int64_t _lastBytes;

  const char* _tmp;
  const char* _tmp2;

  size_t _which;
  std::string _path;

  int64_t _totReadBef;
  int64_t _lastNewData;

  Tag _ret;

  static size_t utf8(size_t cp, char* out);
  const char* emptyStr = "";
};
}

#endif  // XML_FILE_H_