euc_0208

C语言库函数的原型,有用的拿去

// codecvt facet for EUC multibyte code, JIS-X0208 wide-character code
 #pragma once
#ifndef _CVT_EUC_0208_
#define _CVT_EUC_0208_
#ifndef RC_INVOKED
#include <locale>
#include <cwchar>

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

namespace stdext {
	namespace cvt {


typedef _CSTD mbstate_t _Statype;

		// TEMPLATE CLASS codecvt_euc_0208
template<class _Elem,
	unsigned long _Maxcode = 0x7e7e>
	class codecvt_euc_0208
	: public _STD codecvt<_Elem, char, _Statype>
	{	// facet for converting between JIS-X0208 _Elem and EUC bytes
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_euc_0208(size_t _Refs = 0)
		: _Mybase(_Refs)
		{	// construct with ref count
		}

	virtual ~codecvt_euc_0208()
		{	// destroy the object
		}

protected:
	virtual result do_in(_Statype&,
		const _Byte *_First1, const _Byte *_Last1, const _Byte *& _Mid1,
		_Elem *_First2, _Elem *_Last2, _Elem *& _Mid2) const
		{	// convert bytes [_First1, _Last1) to [_First2, _Last)
		_Mid1 = _First1;
		_Mid2 = _First2;

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

			if (_By == 0x8e)
				_Ch = 0, _Extra = true;
			else if (0xa1 <= _By && _By <= 0xfe)
				_Ch = (unsigned short)((_By << 8) - 0x8080), _Extra = true;
			else
				_Ch = _By, _Extra = false;

			if (!_Extra)
				;
			else if (_Last1 - _Mid1 < 2)
				break;
			else if ((_By = (unsigned char)*++_Mid1) < 0xa1 || 0xfe < _By)
				return (_Mybase::error);	// bad second byte
			else
				_Ch = (unsigned short)(_Ch + _By);

			if (_Maxcode < _Ch)
				return (_Mybase::error);	// code too large
			++_Mid1;
			*_Mid2++ = (_Elem)_Ch;
			}

		return (_First1 == _Mid1 ? _Mybase::partial : _Mybase::ok);
		}

	virtual result do_out(_Statype&,
		const _Elem *_First1, const _Elem *_Last1, const _Elem *& _Mid1,
		_Byte *_First2, _Byte *_Last2, _Byte *& _Mid2) const
		{	// convert [_First1, _Last1) to bytes [_First2, _Last)
		_Mid1 = _First1;
		_Mid2 = _First2;

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

			if (_Maxcode < _Ch)
				return (_Mybase::error);

			if (0xa1 <= _Ch && _Ch <= 0xfe)
				if (_Last2 - _Mid2 < 2)
					break;
				else
					{	// put 0x8e plus 1-byte code
					*_Mid2++ = (_Byte)(unsigned short)0x8e;
					*_Mid2++ = (_Byte)_Ch;
					}
			else if (_Ch == 0x8e)
				return (_Mybase::error);
			else if (_Ch <= 0xff)
				*_Mid2++ = (_Byte)_Ch;
			else
				if (_Last2 - _Mid2 < 2)
					break;
				else
					{	// put 2-byte code
					unsigned char _By = (unsigned char)((_Ch >> 8) + 0x80);

					if (_By < 0xa1 || 0xfe < _By)
						return (_Mybase::error);	// bad first byte
					*_Mid2++ = (_Byte)_By;

					_By = (unsigned char)(_Ch + 0x80);
					if (_By < 0xa1 || 0xfe < _By)
						return (_Mybase::error);	// bad second byte
					*_Mid2++ = (_Byte)_By;
					}
			++_Mid1;
			}
		return (_First1 == _Mid1 ? _Mybase::partial : _Mybase::ok);
		}

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

		return (_Mybase::ok);
		}

	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 (3);
		}

	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	0x7e7e
  #define MYFILE	"euc_0208"
  #define MYNAME	stdext::cvt::codecvt_euc_0208<Mywchar>

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

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

#endif /* RC_INVOKED */
#endif /* _CVT_EUC_0208_ */

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