Xport: XHTML Parsing and Objective Reporting Toolkit

/************************************************************************
Xport: XHTML Parsing & Objective Reporting Toolkit
Copyright (C) 2007  Mitchel Haas

This file is part of Xport.

Xport is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.

Xport is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with Xport.  If not, see <http://www.gnu.org/licenses/>.

For complete documentation on this library and alternative
licensing options, visit http://www.xportpro.com
Email questions, comments or suggestions to mitchel.haas@xportpro.com
************************************************************************/
#pragma once

namespace Xport 
{
  template<typename T> class xhtml_entity;
  enum xhtml_entity_appearance {unaltered, character_entity, numeric_entity, named_entity};
  template<typename T> struct entity_ch_comp;
  template<typename T> struct entity_num_comp;
  template<typename T> struct entity_name_comp;
}

template<typename CT>
class Xport::xhtml_entity
{
public:
  xhtml_entity(CT ch, short num, const std::basic_string<CT>& nm, xhtml_entity_appearance ea) : app(ea), character(ch), name(nm), number(num) {}
  // conversion constructors for convenience in set.find() operation
  explicit xhtml_entity(CT ch);
  explicit xhtml_entity(int num) : number(num) {}
  explicit xhtml_entity(const std::basic_string<CT>& nm) : name(nm) {}

  void appearance(xhtml_entity_appearance a) { app = a; }
  xhtml_entity_appearance appearance() const { return app; }
  CT entity_character() const { return character; }
  std::basic_string<CT> entity_name() const { return name; }
  short entity_number() const { return number; }
  static void populate_entities(std::set<xhtml_entity<CT>, entity_num_comp<CT> >&);

private:
  xhtml_entity_appearance app;
  CT character;
  std::basic_string<CT> name;
  short number;

  // friends
  friend struct entity_ch_comp<CT>;
  friend struct entity_num_comp<CT>;
  friend struct entity_name_comp<CT>;

};


template<>
inline Xport::xhtml_entity<char>::xhtml_entity(const char ch) : app(unaltered), character(ch), name(""), number(0)
{

}

template<>
inline Xport::xhtml_entity<wchar_t>::xhtml_entity(const wchar_t ch) : app(unaltered), character(ch), name(L""), number(0)
{
  character <<= 8;
  character >>= 8;
}


template<typename CT>
struct Xport::entity_ch_comp
{
  bool operator() (const xhtml_entity<CT>& lhs, const xhtml_entity<CT>& rhs) const { return lhs.character < rhs.character; }
};

template<typename CT>
struct Xport::entity_num_comp
{
  bool operator() (const xhtml_entity<CT>& lhs, const xhtml_entity<CT>& rhs) const { return lhs.number < rhs.number; }
};

template<typename CT>
struct Xport::entity_name_comp
{
  bool operator() (const xhtml_entity<CT>& lhs, const xhtml_entity<CT>& rhs) const { return lhs.name < rhs.name; }
};

template<typename CT>
inline void Xport::xhtml_entity<CT>::populate_entities(std::set<xhtml_entity<CT>, entity_num_comp<CT> >& entity_vec) 
{
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('"'),   34,   typed_string<CT>("quot"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\''),  39,   typed_string<CT>("apos"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('&'),   38,   typed_string<CT>("amp"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('<'),   60,   typed_string<CT>("lt"),     named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('>'),   62,   typed_string<CT>("gt"),     named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>(' '),   160,  typed_string<CT>("nbsp"),   unaltered));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA1'),  161,  typed_string<CT>("iexcl"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA4'),  164,  typed_string<CT>("curren"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA2'),  162,  typed_string<CT>("cent"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA3'),  163,  typed_string<CT>("pound"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA5'),  165,  typed_string<CT>("yen"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA6'),  166,  typed_string<CT>("brvbar"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA7'),  167,  typed_string<CT>("sect"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA8'),  168,  typed_string<CT>("uml"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xA9'),  169,  typed_string<CT>("copy"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xAA'),  170,  typed_string<CT>("ordf"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xAB'),  171,  typed_string<CT>("laquo"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xAC'),  172,  typed_string<CT>("not"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xAE'),  174,  typed_string<CT>("reg"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xAF'),  175,  typed_string<CT>("macr"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB0'),  176,  typed_string<CT>("deg"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB1'),  177,  typed_string<CT>("plusmn"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB2'),  178,  typed_string<CT>("sup2"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB3'),  179,  typed_string<CT>("sup3"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB4'),  180,  typed_string<CT>("acute"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB5'),  181,  typed_string<CT>("micro"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB6'),  182,  typed_string<CT>("para"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB7'),  183,  typed_string<CT>("middot"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB8'),  184,  typed_string<CT>("cedil"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xB9'),  185,  typed_string<CT>("sup1"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xBA'),  186,  typed_string<CT>("ordm"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xBB'),  187,  typed_string<CT>("raquo"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xBC'),  188,  typed_string<CT>("frac14"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xBD'),  189,  typed_string<CT>("frac12"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xBE'),  190,  typed_string<CT>("frac34"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xBF'),  191,  typed_string<CT>("iquest"),   named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xD7'),  215,  typed_string<CT>("times"),    named_entity));
  entity_vec.insert(xhtml_entity<CT>(typed_char<CT>('\xF7'),  247,  typed_string<CT>("divide"),   named_entity));
}