STL/stl/inc/cvt/xtwo_byte

212 строки
8.7 KiB
C++

// xtwo_byte -- convert one- or two-byte code to 16-bit wide
// Copyright (c) Microsoft Corporation.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
#pragma once
#ifndef _CVT_XTWO_BYTE_
#define _CVT_XTWO_BYTE_
#include <yvals_core.h>
#if _STL_COMPILER_PREPROCESSOR
#include <cwchar>
#include <locale>
#pragma pack(push, _CRT_PACKING)
#pragma warning(push, _STL_WARNING_LEVEL)
#pragma warning(disable : _STL_DISABLED_WARNINGS)
_STL_DISABLE_CLANG_WARNINGS
#pragma push_macro("new")
#undef new
#if !_HAS_IF_CONSTEXPR
#pragma warning(disable : 4127) // conditional expression is constant
#endif // !_HAS_IF_CONSTEXPR
namespace stdext {
namespace cvt {
using _Statype = _CSTD mbstate_t;
// CLASS TEMPLATE _Cvt_two_byte
template <class _Elem, class _Table, unsigned long _Maxcode = 0xffff>
class _Cvt_two_byte
: public _STD
codecvt<_Elem, char, _Statype> { // facet for converting between _Elem and one- or two-byte sequences
public:
using _Mybase = _STD codecvt<_Elem, char, _Statype>;
using result = typename _Mybase::result;
using _Byte = char;
using intern_type = _Elem;
using extern_type = _Byte;
using state_type = _Statype;
explicit _Cvt_two_byte(size_t _Refs = 0) : _Mybase(_Refs) {}
virtual ~_Cvt_two_byte() noexcept {}
protected:
virtual result do_in(_Statype&, const _Byte* _First1, const _Byte* _Last1, const _Byte*& _Mid1,
_Elem* _First2, _Elem* _Last2, _Elem*& _Mid2) const override {
// convert bytes [_First1, _Last1) to [_First2, _Last)
_Mid1 = _First1;
_Mid2 = _First2;
for (; _Mid1 != _Last1 && _Mid2 != _Last2; ++_Mid1) { // get a byte and map it
unsigned char _Bytecode = static_cast<unsigned char>(*_Mid1);
unsigned short _Widecode;
if (_Bytecode < _Table::_Nlow) {
_Widecode = _Bytecode; // map byte to same wide value
} else if ((_Widecode = _Table::_Btw[_Bytecode - _Table::_Nlow]) == 0) {
if _CONSTEXPR_IF (_Table::_Nbytes < 2) {
return _Mybase::error; // no single-byte mapping
} else {
if (_Mid1 + 1 == _Last1) {
break; // need another input byte
}
// form double-byte code and search for a valid mapping
_Widecode =
static_cast<unsigned short>(_Bytecode << 8 | static_cast<unsigned char>(*++_Mid1));
unsigned long _Lo = 0;
unsigned long _Hi = sizeof(_Table::_Dbvalid) / sizeof(_Table::_Dbvalid[0]);
while (_Lo < _Hi) { // test midpoint of remaining interval
unsigned long _Mid = (_Hi + _Lo) / 2;
if (_Widecode < _Table::_Dbvalid[_Mid]) {
_Hi = _Mid;
} else if (_Widecode == _Table::_Dbvalid[_Mid]) { // found the code, map it
_Widecode = _Table::_Dbtw[_Mid];
break;
} else {
_Lo = _Mid + 1;
}
}
if (_Hi <= _Lo) {
return _Mybase::error;
}
}
}
if (_Maxcode < _Widecode) {
return _Mybase::error; // value too large
}
*_Mid2++ = static_cast<_Elem>(_Widecode);
}
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 override {
// convert [_First1, _Last1) to bytes [_First2, _Last)
_Mid1 = _First1;
_Mid2 = _First2;
for (; _Mid1 != _Last1 && _Mid2 != _Last2; ++_Mid1) { // get an element and map it
unsigned long _Hugecode = static_cast<unsigned long>(*_Mid1);
unsigned short _Widecode = static_cast<unsigned short>(_Hugecode);
if (_Maxcode < _Widecode) {
return _Mybase::error; // value too large
}
if (_Widecode >= _Table::_Nlow
&& (_Widecode >= 0x100
|| _Table::_Btw[_Widecode - _Table::_Nlow] != _Widecode)) { // search for a valid mapping
unsigned long _Lo = 0;
unsigned long _Hi = sizeof(_Table::_Wvalid) / sizeof(_Table::_Wvalid[0]);
while (_Lo < _Hi) { // test midpoint of remaining interval
unsigned long _Mid = (_Hi + _Lo) / 2;
if (_Widecode < _Table::_Wvalid[_Mid]) {
_Hi = _Mid;
} else if (_Widecode == _Table::_Wvalid[_Mid]) { // found the code, map it
_Widecode = _Table::_Wtb[_Mid];
break;
} else {
_Lo = _Mid + 1;
}
}
if (_Hi <= _Lo) {
return _Mybase::error;
}
}
if (_Widecode >= 0x100) {
if (_Mid2 + 1 == _Last2) {
return _Mybase::partial; // need room for another byte
}
*_Mid2++ = static_cast<_Byte>(_Widecode >> 8); // put ms byte
}
*_Mid2++ = static_cast<_Byte>(_Widecode);
}
return _First1 == _Mid1 ? _Mybase::partial : _Mybase::ok;
}
virtual result do_unshift(_Statype&, _Byte* _First2, _Byte*, _Byte*& _Mid2) const override {
// generate bytes to return to default shift state
_Mid2 = _First2;
return _Mybase::ok;
}
virtual int do_length(
_Statype& _State, const _Byte* _First1, const _Byte* _Last1, size_t _Count) const noexcept override {
// return min(_Count, converted length of bytes [_First1, _Last1))
size_t _Wchars = 0;
_Statype _Mystate = _State;
while (_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 static_cast<int>(_Wchars + (_Last1 - _First1));
case _Mybase::ok:
if (_Mid2 == &_Ch + 1) {
++_Wchars; // replacement do_in might not convert one
}
_First1 = _Mid1;
break;
default:
return static_cast<int>(_Wchars); // error or partial
}
}
return static_cast<int>(_Wchars);
}
virtual bool do_always_noconv() const noexcept override { // return true if conversions never change input
return false;
}
virtual int do_max_length() const noexcept override { // return maximum length required for a conversion
return _Table::_Nbytes; // length of sequence
}
virtual int do_encoding() const noexcept override { // return length of code sequence (from codecvt)
return 0; // 0 => varying length
}
};
} // namespace cvt
} // namespace stdext
#pragma pop_macro("new")
_STL_RESTORE_CLANG_WARNINGS
#pragma warning(pop)
#pragma pack(pop)
#endif // _STL_COMPILER_PREPROCESSOR
#endif // _CVT_XTWO_BYTE_