xjis

C语言库函数的原型,有用的拿去

// codecvt facet for JIS multibyte code, UCS-2 wide-character code
 #pragma once
#ifndef _CVT_XJIS_
#define _CVT_XJIS_
#ifndef RC_INVOKED
#include <locale>
#include <cwchar>

 #pragma pack(push,_CRT_PACKING)
 #pragma warning(push,3)

 #ifndef _CVT_JIS_
  #error do not include directly
 #endif /* _CVT_JIS */

namespace stdext {
	namespace cvt {

typedef _CSTD mbstate_t _Statype;

 #define _ESC_CODE	0x1b

		// TEMPLATE CLASS codecvt_jis
template<class _Elem,
	unsigned long _Maxcode = 0xffff>
	class codecvt_jis
	: public _STD codecvt<_Elem, char, _Statype>
	{	// facet for converting between JIS-X0208 _Elem and UCS-2 bytes
	typedef _tab_jis<int> _Table;

	bool _Jis_to_ucs(unsigned short& _Ch) const
		{	// convert _Ch from JIS-X0208 to UCS-2
		if (_Ch < _Table::_Nlow)
			;	// map byte to same wide value
		else if (_Ch < 0x100)
			{	// map a one-byte code
			if ((_Ch = _Table::_Btw[_Ch - _Table::_Nlow]) == 0)
				return (false);	// no defined mapping for byte
			}
		else
			{	// search for a valid mapping
			unsigned long _Lo = 0;
			unsigned long _Hi = sizeof (_Table::_Dbvalid)
				/ sizeof (_Table::_Dbvalid[0]);

			for (; _Lo < _Hi; )
				{	// test midpoint of remaining interval
				unsigned long _Mid = (_Hi + _Lo) / 2;

				if (_Ch < _Table::_Dbvalid[_Mid])
					_Hi = _Mid;
				else if (_Ch == _Table::_Dbvalid[_Mid])
					{	// found the code, map it
					_Ch = _Table::_Dbtw[_Mid];
					return (true);
					}
				else
					_Lo = _Mid + 1;
				}
			return (false);
			}
		return (true);
		}

	bool _Ucs_to_jis(unsigned short& _Ch) const
		{	// convert _Ch from UCS-2 to JIS-X0208
		if (_Ch < _Table::_Nlow || _Ch < 0x100
			&& _Table::_Btw[_Ch - _Table::_Nlow] == _Ch)
			;	// map wide to same byte value
		else
			{	// search for a valid mapping
			unsigned long _Lo = 0;
			unsigned long _Hi = sizeof (_Table::_Wvalid)
				/ sizeof (_Table::_Wvalid[0]);

			for (; _Lo < _Hi; )
				{	// test midpoint of remaining interval
				unsigned long _Mid = (_Hi + _Lo) / 2;

				if (_Ch < _Table::_Wvalid[_Mid])
					_Hi = _Mid;
				else if (_Ch == _Table::_Wvalid[_Mid])
					{	// found the code, map it
					_Ch = _Table::_Wtb[_Mid];
					return (true);
					}
				else
					_Lo = _Mid + 1;
				}
			return (false);
			}
		return (true);
		}

public:
 	typedef _STD codecvt<_Elem, char, _Statype> _Mybase;
	typedef typename _Mybase::result result;
	typedef char _Byte;
	typedef _Elem intern_type;
	typedef _Byte extern_type;
	typedef _Statype state_type;

	explicit codecvt_jis(size_t _Refs = 0)
		: _Mybase(_Refs)
		{	// construct with ref count
		}

	virtual ~codecvt_jis()
		{	// destroy the object
		}

protected:
	virtual result do_in(_Statype& _State,
		const _Byte *_First1, const _Byte *_Last1, const _Byte *& _Mid1,
		_Elem *_First2, _Elem *_Last2, _Elem *& _Mid2) const
		{	// convert bytes [_First1, _Last1) to [_First2, _Last)
		unsigned long _Maxc = _Maxcode;	// to quiet warnings
		char *_Pstate = (char *)&_State;
		result _Ans = _Mybase::partial;

		_Mid1 = _First1;
		_Mid2 = _First2;
		for (; _Mid1 != _Last1 && _Mid2 != _Last2; )
			{	// convert a multibyte sequence
			unsigned char _By = (unsigned char)*_Mid1;
			unsigned short _Ch;

			if (_By != _ESC_CODE)
				;
			else if (_Last1 - _Mid1 < 4)
				break;
			else if (_Mid1[1] == '('
				&& (_Mid1[2] == 'B' || _Mid1[2] == 'J')
				&& _Mid1[3] != _ESC_CODE)
				{	// shift to 1-byte mode
				_By = *(_Mid1 += 3);
				*_Pstate = 0;
				continue;
				}
			else if (_Mid1[1] == '$'
				&& (_Mid1[2] == 'B' || _Mid1[2] == '@'))
				if (_Last1 - _Mid1 < 5)
					break;	// not enough bytes for escape plus 2-byte code
				else
					{	// enough bytes, shift and convert
				_By = *(_Mid1 += 3);
					*_Pstate = 1;
					continue;
					}
			else
				return (_Mybase::error);	// bad escape sequence

			_Ch = _By;
			if (*_Pstate == 0)
				;
			else if (_By < 0x21 || 0x7e < _By)
				return (_Mybase::error);	// bad first byte
			else if (_Last1 - _Mid1 < 2)
				break;
			else if ((_By = *++_Mid1) < 0x21 || 0x7e < _By)
				return (_Mybase::error);	// bad second byte
			else
				_Ch = (unsigned short)(_Ch << 8 | _By);

			if (!_Jis_to_ucs(_Ch) || _Maxc < _Ch)
				return (_Mybase::error);	// code too large
			++_Mid1;
			*_Mid2++ = (_Elem)_Ch;
			_Ans = _Mybase::ok;
			}

		return (_Ans);
		}

	virtual result do_out(_Statype& _State,
		const _Elem *_First1, const _Elem *_Last1, const _Elem *& _Mid1,
		_Byte *_First2, _Byte *_Last2, _Byte *& _Mid2) const
		{	// convert [_First1, _Last1) to bytes [_First2, _Last)
		unsigned long _Maxc = _Maxcode;	// to quiet warnings
		char *_Pstate = (char *)&_State;

		_Mid1 = _First1;
		_Mid2 = _First2;
		for (; _Mid1 != _Last1 && _Mid2 != _Last2; )
			{	// convert and put a wide char
			unsigned long _Uch = (unsigned long)*_Mid1;
			unsigned short _Ch = (unsigned short)_Uch;

			if (_Maxc < _Ch || !_Ucs_to_jis(_Ch))
				return (_Mybase::error);

			if (_Ch <= 0xff)
				{	// put a 1-byte code
				if (_Ch == _ESC_CODE)
					return (_Mybase::error);	// can't output bald ESC
				else if (*_Pstate == 0)
					;	// already in proper state
				else if (_Last2 - _Mid2 < 4)
					break;
				else
					{	// put shift to 1-byte state
					*_Mid2++ = _ESC_CODE;
					*_Mid2++ = '(';
					*_Mid2++ = 'B';
					*_Pstate = 0;
					}
				*_Mid2++ = (_Byte)_Ch;
				}
			else
				{	// put a 2-byte code
				if (*_Pstate != 0)
					{	// already in proper state, check room
					if (_Last2 - _Mid2 < 2)
						break;
					}
				else if (_Last2 - _Mid2 < 5)
					break;
				else
					{	// put shift to 2-byte state
					*_Mid2++ = _ESC_CODE;
					*_Mid2++ = '$';
					*_Mid2++ = 'B';
					*_Pstate = 1;
					}

				unsigned char _By = (unsigned char)(_Ch >> 8);

				if (_By < 0x21 || 0x7e < _By)
					return (_Mybase::error);	// bad first byte
				*_Mid2++ = (_Byte)_By;

				_By = (unsigned char)_Ch;
				if (_By < 0x21 || 0x7e < _By)
					return (_Mybase::error);	// bad first byte
				*_Mid2++ = (_Byte)_By;
				}
			++_Mid1;
			}
		return (_First1 == _Mid1 ? _Mybase::partial : _Mybase::ok);
		}

	virtual result do_unshift(_Statype& _State,
		_Byte *_First2, _Byte *, _Byte *& _Mid2) const
		{	// generate bytes to return to default shift state
		char *_Pstate = (char *)&_State;

		_Mid2 = _First2;
		switch (*_Pstate)
			{
		case 1:	// need to home, put first of three bytes
			*_Pstate = 2;
			*_Mid2++ = _ESC_CODE;
			break;

		case 2:	// put second of three bytes
			*_Pstate = 3;
			*_Mid2++ = '(';
			break;

		case 3:	// put third of three bytes
			*_Pstate = 0;
			*_Mid2++ = 'B';

		case 0:	// fall through
			return (_Mybase::ok);
			}
		return (_Mybase::partial);
		}

	virtual int do_length(const _Statype& _State, const _Byte *_First1,
		const _Byte *_Last1, size_t _Count) const _THROW0()
		{	// return min(_Count, converted length of bytes [_First1, _Last1))
		size_t _Wchars = 0;
		_Statype _Mystate = _State;

		for (; _Wchars < _Count && _First1 != _Last1; )
			{	// convert another wide character
			const _Byte *_Mid1;
			_Elem *_Mid2;
			_Elem _Ch;

			switch (do_in(_Mystate, _First1, _Last1, _Mid1,
				&_Ch, &_Ch + 1, _Mid2))
				{	// test result of single wide-char conversion
			case _Mybase::noconv:
				return ((int)(_Wchars + (_Last1 - _First1)));

			case  _Mybase::ok:
				if (_Mid2 == &_Ch + 1)
					++_Wchars;	// replacement do_in might not convert one
				_First1 = _Mid1;
				break;

			default:
				return ((int)_Wchars);	// error or partial
				}
			}

		return ((int)_Wchars);
		}

	virtual bool do_always_noconv() const _THROW0()
		{	// return true if conversions never change input
		return (false);
		}

	virtual int do_max_length() const _THROW0()
		{	// return maximum length required for a conversion
		return (5);
		}

	virtual int do_encoding() const _THROW0()
		{	// return length of code sequence (from codecvt)
		return (0);	// 0 => varying length
		}
	};
		}	// namespace cvt
	}	// namespace stdext

 #ifdef _TEST_IT
  #define NCHARS	0x10000
  #define MYWC_MAX	0xffff
  #define MYFILE	"xjis"
  #define MYNAME	stdext::cvt::codecvt_jis<Mywchar>

  #include <cvt/xtest>
 #endif /* _TEST_IT */

 #pragma warning(pop)
 #pragma pack(pop)

#endif /* RC_INVOKED */
#endif /* _CVT_XJIS_ */

/*
 * Copyright (c) 1992-2009 by P.J. Plauger.  ALL RIGHTS RESERVED. 
 * Consult your license regarding permissions and restrictions.
V5.20:0009 */