зеркало из https://github.com/microsoft/STL.git
693 строки
26 KiB
C++
693 строки
26 KiB
C++
// codecvt standard header
|
|
|
|
// Copyright (c) Microsoft Corporation.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
|
|
#ifndef _CODECVT_
|
|
#define _CODECVT_
|
|
#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
|
|
|
|
_STD_BEGIN
|
|
_INLINE_VAR constexpr int _Little_first = 1;
|
|
_INLINE_VAR constexpr int _Big_first = 2;
|
|
|
|
_EXPORT_STD enum _CXX17_DEPRECATE_CODECVT_HEADER codecvt_mode {
|
|
consume_header = 4,
|
|
generate_header = 2,
|
|
little_endian = 1
|
|
};
|
|
|
|
using _Statype = _CSTD mbstate_t;
|
|
|
|
_STL_DISABLE_DEPRECATED_WARNING
|
|
_EXPORT_STD template <class _Elem, unsigned long _Mymax = 0x10ffff, codecvt_mode _Mymode = codecvt_mode{}>
|
|
class _CXX17_DEPRECATE_CODECVT_HEADER codecvt_utf8 : public codecvt<_Elem, char, _Statype> {
|
|
// facet for converting between _Elem and UTF-8 byte sequences
|
|
public:
|
|
using _Mybase = 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 codecvt_utf8(size_t _Refs = 0) : _Mybase(_Refs) {}
|
|
|
|
__CLR_OR_THIS_CALL ~codecvt_utf8() noexcept override {}
|
|
|
|
protected:
|
|
result __CLR_OR_THIS_CALL do_in(_Statype& _State, const _Byte* _First1, const _Byte* _Last1, const _Byte*& _Mid1,
|
|
_Elem* _First2, _Elem* _Last2, _Elem*& _Mid2) const override {
|
|
// convert bytes [_First1, _Last1) to [_First2, _Last2)
|
|
char* _Pstate = reinterpret_cast<char*>(&_State);
|
|
_Mid1 = _First1;
|
|
_Mid2 = _First2;
|
|
|
|
while (_Mid1 != _Last1 && _Mid2 != _Last2) { // convert a multibyte sequence
|
|
unsigned long _By = static_cast<unsigned char>(*_Mid1);
|
|
unsigned long _Ch;
|
|
int _Nextra;
|
|
|
|
if (_By < 0x80u) {
|
|
_Ch = _By;
|
|
_Nextra = 0;
|
|
} else if (_By < 0xc0u) { // 0x80-0xbf not first byte
|
|
++_Mid1;
|
|
return _Mybase::error;
|
|
} else if (_By < 0xe0u) {
|
|
_Ch = _By & 0x1f;
|
|
_Nextra = 1;
|
|
} else if (_By < 0xf0u) {
|
|
_Ch = _By & 0x0f;
|
|
_Nextra = 2;
|
|
} else if (_By < 0xf8u) {
|
|
_Ch = _By & 0x07;
|
|
_Nextra = 3;
|
|
} else {
|
|
_Ch = _By & 0x03;
|
|
_Nextra = _By < 0xfc ? 4 : 5;
|
|
}
|
|
|
|
if (_Nextra == 0) {
|
|
++_Mid1;
|
|
} else if (_Last1 - _Mid1 < _Nextra + 1) {
|
|
break; // not enough input
|
|
} else {
|
|
for (++_Mid1; 0 < _Nextra; --_Nextra, ++_Mid1) {
|
|
if ((_By = static_cast<unsigned char>(*_Mid1)) < 0x80u || 0xc0u <= _By) {
|
|
return _Mybase::error; // not continuation byte
|
|
} else {
|
|
_Ch = _Ch << 6 | (_By & 0x3f);
|
|
}
|
|
}
|
|
}
|
|
|
|
if (*_Pstate == 0) { // first time, maybe look for and consume header
|
|
*_Pstate = 1;
|
|
|
|
constexpr bool _Consuming = (_Mymode & consume_header) != 0;
|
|
if constexpr (_Consuming) {
|
|
if (_Ch == 0xfeff) { // drop header and retry
|
|
const result _Ans = do_in(_State, _Mid1, _Last1, _Mid1, _First2, _Last2, _Mid2);
|
|
|
|
if (_Ans == _Mybase::partial) { // roll back header determination
|
|
*_Pstate = 0;
|
|
_Mid1 = _First1;
|
|
}
|
|
|
|
return _Ans;
|
|
}
|
|
}
|
|
}
|
|
|
|
if (_Mymax < _Ch) {
|
|
return _Mybase::error; // code too large
|
|
}
|
|
|
|
*_Mid2++ = static_cast<_Elem>(_Ch);
|
|
}
|
|
|
|
return _First1 == _Mid1 ? _Mybase::partial : _Mybase::ok;
|
|
}
|
|
|
|
result __CLR_OR_THIS_CALL do_out(_Statype& _State, const _Elem* _First1, const _Elem* _Last1, const _Elem*& _Mid1,
|
|
_Byte* _First2, _Byte* _Last2, _Byte*& _Mid2) const override {
|
|
// convert [_First1, _Last1) to bytes [_First2, _Last2)
|
|
char* _Pstate = reinterpret_cast<char*>(&_State);
|
|
_Mid1 = _First1;
|
|
_Mid2 = _First2;
|
|
|
|
while (_Mid1 != _Last1 && _Mid2 != _Last2) { // convert and put a wide char
|
|
_Byte _By;
|
|
int _Nextra;
|
|
unsigned long _Ch = static_cast<unsigned long>(*_Mid1);
|
|
|
|
if (_Mymax < _Ch) {
|
|
return _Mybase::error;
|
|
}
|
|
|
|
if (_Ch < 0x0080u) {
|
|
_By = static_cast<_Byte>(_Ch);
|
|
_Nextra = 0;
|
|
} else if (_Ch < 0x0800u) {
|
|
_By = static_cast<_Byte>(0xc0 | _Ch >> 6);
|
|
_Nextra = 1;
|
|
} else if (_Ch < 0x00010000u) {
|
|
_By = static_cast<_Byte>(0xe0 | _Ch >> 12);
|
|
_Nextra = 2;
|
|
} else if (_Ch < 0x00200000u) {
|
|
_By = static_cast<_Byte>(0xf0 | _Ch >> 18);
|
|
_Nextra = 3;
|
|
} else if (_Ch < 0x04000000u) {
|
|
_By = static_cast<_Byte>(0xf8 | _Ch >> 24);
|
|
_Nextra = 4;
|
|
} else {
|
|
_By = static_cast<_Byte>(0xfc | (_Ch >> 30 & 0x03));
|
|
_Nextra = 5;
|
|
}
|
|
|
|
if (*_Pstate == 0) { // first time, maybe generate header
|
|
*_Pstate = 1;
|
|
constexpr bool _Generating = (_Mymode & generate_header) != 0;
|
|
if constexpr (_Generating) {
|
|
if (_Last2 - _Mid2 < 3 + 1 + _Nextra) {
|
|
return _Mybase::partial; // not enough room for both
|
|
}
|
|
|
|
// prepend header
|
|
*_Mid2++ = '\xef';
|
|
*_Mid2++ = '\xbb';
|
|
*_Mid2++ = '\xbf';
|
|
}
|
|
}
|
|
|
|
if (_Last2 - _Mid2 < 1 + _Nextra) {
|
|
break; // not enough room for output
|
|
}
|
|
|
|
++_Mid1;
|
|
for (*_Mid2++ = _By; 0 < _Nextra;) {
|
|
*_Mid2++ = static_cast<_Byte>((_Ch >> 6 * --_Nextra & 0x3f) | 0x80);
|
|
}
|
|
}
|
|
|
|
return _First1 == _Mid1 ? _Mybase::partial : _Mybase::ok;
|
|
}
|
|
|
|
result __CLR_OR_THIS_CALL do_unshift(_Statype&, _Byte* _First2, _Byte*, _Byte*& _Mid2) const override {
|
|
// generate bytes to return to default shift state
|
|
_Mid2 = _First2;
|
|
return _Mybase::noconv;
|
|
}
|
|
|
|
friend int _Codecvt_do_length<>(const codecvt_utf8&, _Statype&, const _Byte*, const _Byte*, size_t);
|
|
|
|
int __CLR_OR_THIS_CALL do_length(
|
|
_Statype& _State, const _Byte* _First1, const _Byte* _Last1, size_t _Count) const noexcept override {
|
|
return _Codecvt_do_length(*this, _State, _First1, _Last1, _Count);
|
|
}
|
|
|
|
bool __CLR_OR_THIS_CALL do_always_noconv() const noexcept override {
|
|
// return true if conversions never change input
|
|
return false;
|
|
}
|
|
|
|
int __CLR_OR_THIS_CALL do_max_length() const noexcept override {
|
|
// return maximum length required for a conversion
|
|
if constexpr ((_Mymode & (consume_header | generate_header)) != 0) {
|
|
return 9;
|
|
} else {
|
|
return 6;
|
|
}
|
|
}
|
|
|
|
int __CLR_OR_THIS_CALL do_encoding() const noexcept override {
|
|
// return length of code sequence (from codecvt)
|
|
if constexpr ((_Mymode & (consume_header | generate_header)) != 0) {
|
|
return -1; // -1 => state dependent
|
|
} else {
|
|
return 0; // 0 => varying length
|
|
}
|
|
}
|
|
};
|
|
|
|
_EXPORT_STD template <class _Elem, unsigned long _Mymax = 0x10ffff, codecvt_mode _Mymode = codecvt_mode{}>
|
|
class _CXX17_DEPRECATE_CODECVT_HEADER codecvt_utf16 : public codecvt<_Elem, char, _Statype> {
|
|
// facet for converting between _Elem and UTF-16 multibyte sequences
|
|
private:
|
|
enum { _Bytes_per_word = 2 };
|
|
|
|
public:
|
|
using _Mybase = 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 codecvt_utf16(size_t _Refs = 0) : _Mybase(_Refs) {}
|
|
|
|
__CLR_OR_THIS_CALL ~codecvt_utf16() noexcept override {}
|
|
|
|
protected:
|
|
result __CLR_OR_THIS_CALL do_in(_Statype& _State, const _Byte* _First1, const _Byte* _Last1, const _Byte*& _Mid1,
|
|
_Elem* _First2, _Elem* _Last2, _Elem*& _Mid2) const override {
|
|
// convert bytes [_First1, _Last1) to [_First2, _Last2)
|
|
char* _Pstate = reinterpret_cast<char*>(&_State);
|
|
_Mid1 = _First1;
|
|
_Mid2 = _First2;
|
|
|
|
while (_Bytes_per_word <= _Last1 - _Mid1 && _Mid2 != _Last2) { // convert a multibyte sequence
|
|
const auto _Ptr = reinterpret_cast<const unsigned char*>(_Mid1);
|
|
unsigned long _Ch;
|
|
unsigned short _Ch0;
|
|
unsigned short _Ch1;
|
|
|
|
if (*_Pstate == _Little_first) {
|
|
_Ch0 = static_cast<unsigned short>(_Ptr[1] << 8 | _Ptr[0]);
|
|
} else if (*_Pstate == _Big_first) {
|
|
_Ch0 = static_cast<unsigned short>(_Ptr[0] << 8 | _Ptr[1]);
|
|
} else { // no header seen yet, try preferred mode
|
|
constexpr bool _Prefer_LE = (_Mymode & little_endian) != 0;
|
|
constexpr char _Default_endian = _Prefer_LE ? _Little_first : _Big_first;
|
|
|
|
if constexpr (_Prefer_LE) {
|
|
_Ch0 = static_cast<unsigned short>(_Ptr[1] << 8 | _Ptr[0]);
|
|
} else {
|
|
_Ch0 = static_cast<unsigned short>(_Ptr[0] << 8 | _Ptr[1]);
|
|
}
|
|
|
|
*_Pstate = _Default_endian;
|
|
constexpr bool _Consuming = (_Mymode & consume_header) != 0;
|
|
if constexpr (_Consuming) {
|
|
if (_Ch0 == 0xfffeu) {
|
|
*_Pstate = 3 - _Default_endian;
|
|
}
|
|
|
|
if (_Ch0 == 0xfffeu || _Ch0 == 0xfeffu) { // consume header, fixate on endianness, and retry
|
|
_Mid1 += _Bytes_per_word;
|
|
result _Ans = do_in(_State, _Mid1, _Last1, _Mid1, _First2, _Last2, _Mid2);
|
|
|
|
if (_Ans == _Mybase::partial) { // not enough bytes, roll back header
|
|
*_Pstate = 0;
|
|
_Mid1 = _First1;
|
|
}
|
|
|
|
return _Ans;
|
|
}
|
|
}
|
|
}
|
|
|
|
if (_Ch0 < 0xd800u || 0xdc00u <= _Ch0) { // one word, consume bytes
|
|
_Mid1 += _Bytes_per_word;
|
|
_Ch = _Ch0;
|
|
} else if (_Last1 - _Mid1 < 2 * _Bytes_per_word) {
|
|
break;
|
|
} else { // get second word
|
|
if (*_Pstate == _Little_first) {
|
|
_Ch1 = static_cast<unsigned short>(_Ptr[3] << 8 | _Ptr[2]);
|
|
} else {
|
|
_Ch1 = static_cast<unsigned short>(_Ptr[2] << 8 | _Ptr[3]);
|
|
}
|
|
|
|
if (_Ch1 < 0xdc00u || 0xe000u <= _Ch1) {
|
|
return _Mybase::error;
|
|
}
|
|
|
|
_Mid1 += 2 * _Bytes_per_word;
|
|
_Ch = static_cast<unsigned long>(_Ch0 - 0xd800 + 0x0040) << 10 | (_Ch1 - 0xdc00);
|
|
}
|
|
|
|
if (_Mymax < _Ch) {
|
|
return _Mybase::error; // code too large
|
|
}
|
|
|
|
*_Mid2++ = static_cast<_Elem>(_Ch);
|
|
}
|
|
|
|
return _First1 == _Mid1 ? _Mybase::partial : _Mybase::ok;
|
|
}
|
|
|
|
result __CLR_OR_THIS_CALL do_out(_Statype& _State, const _Elem* _First1, const _Elem* _Last1, const _Elem*& _Mid1,
|
|
_Byte* _First2, _Byte* _Last2, _Byte*& _Mid2) const override {
|
|
// convert [_First1, _Last1) to bytes [_First2, _Last2)
|
|
char* _Pstate = reinterpret_cast<char*>(&_State);
|
|
_Mid1 = _First1;
|
|
_Mid2 = _First2;
|
|
|
|
if (*_Pstate == 0) { // determine endianness once, maybe generate header
|
|
if constexpr ((_Mymode & little_endian) != 0) {
|
|
*_Pstate = _Little_first;
|
|
} else {
|
|
*_Pstate = _Big_first;
|
|
}
|
|
|
|
constexpr bool _Generating = (_Mymode & generate_header) != 0;
|
|
if constexpr (_Generating) {
|
|
if (_Last2 - _Mid2 < 3 * _Bytes_per_word) {
|
|
return _Mybase::partial; // not enough room for all
|
|
}
|
|
|
|
if (*_Pstate == _Little_first) { // put header LS byte first
|
|
*_Mid2++ = '\xff';
|
|
*_Mid2++ = '\xfe';
|
|
} else { // put header MS byte first
|
|
*_Mid2++ = '\xfe';
|
|
*_Mid2++ = '\xff';
|
|
}
|
|
}
|
|
}
|
|
|
|
while (_Mid1 != _Last1 && _Bytes_per_word <= _Last2 - _Mid2) { // convert and put a wide char
|
|
bool _Extra = false;
|
|
unsigned long _Ch = static_cast<unsigned long>(*_Mid1++);
|
|
|
|
if ((_Mymax < 0x10ffffu ? _Mymax : 0x10ffffu) < _Ch) {
|
|
return _Mybase::error; // value too large
|
|
}
|
|
|
|
if (_Ch <= 0xffffu) { // one word, can't be code for first of two
|
|
if (0xd800u <= _Ch && _Ch < 0xdc00u) {
|
|
return _Mybase::error;
|
|
}
|
|
} else if (_Last2 - _Mid2 < 2 * _Bytes_per_word) { // not enough room for two-word output, back up
|
|
--_Mid1;
|
|
return _Mybase::partial;
|
|
} else {
|
|
_Extra = true;
|
|
}
|
|
|
|
if (*_Pstate == _Little_first) {
|
|
if (_Extra) { // put a pair of words LS byte first
|
|
unsigned short _Ch0 =
|
|
static_cast<unsigned short>(0xd800 | static_cast<unsigned short>(_Ch >> 10) - 0x0040);
|
|
*_Mid2++ = static_cast<_Byte>(_Ch0);
|
|
*_Mid2++ = static_cast<_Byte>(_Ch0 >> 8);
|
|
|
|
_Ch0 = static_cast<unsigned short>(0xdc00 | (static_cast<unsigned short>(_Ch) & 0x03ff));
|
|
*_Mid2++ = static_cast<_Byte>(_Ch0);
|
|
*_Mid2++ = static_cast<_Byte>(_Ch0 >> 8);
|
|
} else { // put a single word LS byte first
|
|
*_Mid2++ = static_cast<_Byte>(_Ch);
|
|
*_Mid2++ = static_cast<_Byte>(_Ch >> 8);
|
|
}
|
|
} else {
|
|
if (_Extra) { // put a pair of words MS byte first
|
|
unsigned short _Ch0 =
|
|
static_cast<unsigned short>(0xd800 | static_cast<unsigned short>(_Ch >> 10) - 0x0040);
|
|
*_Mid2++ = static_cast<_Byte>(_Ch0 >> 8);
|
|
*_Mid2++ = static_cast<_Byte>(_Ch0);
|
|
|
|
_Ch0 = static_cast<unsigned short>(0xdc00 | (static_cast<unsigned short>(_Ch) & 0x03ff));
|
|
*_Mid2++ = static_cast<_Byte>(_Ch0 >> 8);
|
|
*_Mid2++ = static_cast<_Byte>(_Ch0);
|
|
} else { // put a single word MS byte first
|
|
*_Mid2++ = static_cast<_Byte>(_Ch >> 8);
|
|
*_Mid2++ = static_cast<_Byte>(_Ch);
|
|
}
|
|
}
|
|
}
|
|
|
|
return _First1 == _Mid1 ? _Mybase::partial : _Mybase::ok;
|
|
}
|
|
|
|
result __CLR_OR_THIS_CALL do_unshift(_Statype&, _Byte* _First2, _Byte*, _Byte*& _Mid2) const override {
|
|
// generate bytes to return to default shift state
|
|
_Mid2 = _First2;
|
|
return _Mybase::noconv;
|
|
}
|
|
|
|
friend int _Codecvt_do_length<>(const codecvt_utf16&, _Statype&, const _Byte*, const _Byte*, size_t);
|
|
|
|
int __CLR_OR_THIS_CALL do_length(
|
|
_Statype& _State, const _Byte* _First1, const _Byte* _Last1, size_t _Count) const noexcept override {
|
|
return _Codecvt_do_length(*this, _State, _First1, _Last1, _Count);
|
|
}
|
|
|
|
bool __CLR_OR_THIS_CALL do_always_noconv() const noexcept override {
|
|
// return true if conversions never change input
|
|
return false;
|
|
}
|
|
|
|
int __CLR_OR_THIS_CALL do_max_length() const noexcept override {
|
|
// return maximum length required for a conversion
|
|
if constexpr ((_Mymode & (consume_header | generate_header)) != 0) {
|
|
return 3 * _Bytes_per_word;
|
|
} else {
|
|
return 6 * _Bytes_per_word;
|
|
}
|
|
}
|
|
|
|
int __CLR_OR_THIS_CALL do_encoding() const noexcept override {
|
|
// return length of code sequence (from codecvt)
|
|
if constexpr ((_Mymode & (consume_header | generate_header)) != 0) {
|
|
return -1; // -1 => state dependent
|
|
} else {
|
|
return 0; // 0 => varying length
|
|
}
|
|
}
|
|
};
|
|
|
|
_EXPORT_STD template <class _Elem, unsigned long _Mymax = 0x10ffff, codecvt_mode _Mymode = codecvt_mode{}>
|
|
class _CXX17_DEPRECATE_CODECVT_HEADER codecvt_utf8_utf16
|
|
: public codecvt<_Elem, char, _Statype> { // facet for converting between UTF-16 _Elem and UTF-8 byte sequences
|
|
public:
|
|
using _Mybase = codecvt<_Elem, char, _Statype>;
|
|
using result = typename _Mybase::result;
|
|
using _Byte = char;
|
|
using intern_type = _Elem;
|
|
using extern_type = _Byte;
|
|
using state_type = _Statype;
|
|
|
|
static_assert(sizeof(unsigned short) <= sizeof(state_type), "state_type too small");
|
|
|
|
explicit codecvt_utf8_utf16(size_t _Refs = 0) : _Mybase(_Refs) {}
|
|
|
|
__CLR_OR_THIS_CALL ~codecvt_utf8_utf16() noexcept override {}
|
|
|
|
protected:
|
|
result __CLR_OR_THIS_CALL do_in(_Statype& _State, const _Byte* _First1, const _Byte* _Last1, const _Byte*& _Mid1,
|
|
_Elem* _First2, _Elem* _Last2, _Elem*& _Mid2) const override {
|
|
// convert bytes [_First1, _Last1) to [_First2, _Last2)
|
|
unsigned short* _Pstate = reinterpret_cast<unsigned short*>(&_State);
|
|
_Mid1 = _First1;
|
|
_Mid2 = _First2;
|
|
|
|
while (_Mid1 != _Last1 && _Mid2 != _Last2) { // convert a multibyte sequence
|
|
unsigned long _By = static_cast<unsigned char>(*_Mid1);
|
|
unsigned long _Ch;
|
|
int _Nextra;
|
|
int _Nskip;
|
|
|
|
if (*_Pstate > 1u) {
|
|
if (_By < 0x80u || 0xc0u <= _By) {
|
|
return _Mybase::error; // not continuation byte
|
|
}
|
|
|
|
// deliver second half of two-word value
|
|
++_Mid1;
|
|
*_Mid2++ = static_cast<_Elem>(*_Pstate | (_By & 0x3f));
|
|
*_Pstate = 1;
|
|
continue;
|
|
}
|
|
|
|
if (_By < 0x80u) {
|
|
_Ch = _By;
|
|
_Nextra = 0;
|
|
} else if (_By < 0xc0u) { // 0x80-0xbf not first byte
|
|
++_Mid1;
|
|
return _Mybase::error;
|
|
} else if (_By < 0xe0u) {
|
|
_Ch = _By & 0x1f;
|
|
_Nextra = 1;
|
|
} else if (_By < 0xf0u) {
|
|
_Ch = _By & 0x0f;
|
|
_Nextra = 2;
|
|
} else if (_By < 0xf8u) {
|
|
_Ch = _By & 0x07;
|
|
_Nextra = 3;
|
|
} else {
|
|
_Ch = _By & 0x03;
|
|
_Nextra = _By < 0xfc ? 4 : 5;
|
|
}
|
|
|
|
_Nskip = _Nextra < 3 ? 0 : 1; // leave a byte for 2nd word
|
|
_First1 = _Mid1; // roll back point
|
|
|
|
if (_Nextra == 0) {
|
|
++_Mid1;
|
|
} else if (_Last1 - _Mid1 < _Nextra + 1 - _Nskip) {
|
|
break; // not enough input
|
|
} else {
|
|
for (++_Mid1; _Nskip < _Nextra; --_Nextra, ++_Mid1) {
|
|
if ((_By = static_cast<unsigned char>(*_Mid1)) < 0x80u || 0xc0u <= _By) {
|
|
return _Mybase::error; // not continuation byte
|
|
}
|
|
|
|
_Ch = _Ch << 6 | (_By & 0x3f);
|
|
}
|
|
}
|
|
|
|
if (0 < _Nskip) {
|
|
_Ch <<= 6; // get last byte on next call
|
|
}
|
|
|
|
if ((_Mymax < 0x10ffffu ? _Mymax : 0x10ffffu) < _Ch) {
|
|
return _Mybase::error; // value too large
|
|
}
|
|
|
|
if (0xffffu < _Ch) { // deliver first half of two-word value, save second word
|
|
unsigned short _Ch0 = static_cast<unsigned short>(0xd800 | (_Ch >> 10) - 0x0040);
|
|
|
|
*_Mid2++ = static_cast<_Elem>(_Ch0);
|
|
*_Pstate = static_cast<unsigned short>(0xdc00 | (_Ch & 0x03ff));
|
|
continue;
|
|
}
|
|
|
|
if (_Nskip != 0) {
|
|
if (_Mid1 == _Last1) { // not enough bytes, noncanonical value
|
|
_Mid1 = _First1;
|
|
break;
|
|
}
|
|
|
|
if ((_By = static_cast<unsigned char>(*_Mid1++)) < 0x80u || 0xc0u <= _By) {
|
|
return _Mybase::error; // not continuation byte
|
|
}
|
|
|
|
_Ch |= _By & 0x3f; // complete noncanonical value
|
|
}
|
|
|
|
if (*_Pstate == 0u) { // first time, maybe look for and consume header
|
|
*_Pstate = 1;
|
|
|
|
constexpr bool _Consuming = (_Mymode & consume_header) != 0;
|
|
if constexpr (_Consuming) {
|
|
if (_Ch == 0xfeffu) { // drop header and retry
|
|
result _Ans = do_in(_State, _Mid1, _Last1, _Mid1, _First2, _Last2, _Mid2);
|
|
|
|
if (_Ans == _Mybase::partial) { // roll back header determination
|
|
*_Pstate = 0;
|
|
_Mid1 = _First1;
|
|
}
|
|
|
|
return _Ans;
|
|
}
|
|
}
|
|
}
|
|
|
|
*_Mid2++ = static_cast<_Elem>(_Ch);
|
|
}
|
|
|
|
return _First1 == _Mid1 ? _Mybase::partial : _Mybase::ok;
|
|
}
|
|
|
|
result __CLR_OR_THIS_CALL do_out(_Statype& _State, const _Elem* _First1, const _Elem* _Last1, const _Elem*& _Mid1,
|
|
_Byte* _First2, _Byte* _Last2, _Byte*& _Mid2) const override {
|
|
// convert [_First1, _Last1) to bytes [_First2, _Last2)
|
|
unsigned short* _Pstate = reinterpret_cast<unsigned short*>(&_State);
|
|
_Mid1 = _First1;
|
|
_Mid2 = _First2;
|
|
|
|
while (_Mid1 != _Last1 && _Mid2 != _Last2) { // convert and put a wide char
|
|
unsigned long _Ch;
|
|
unsigned short _Ch1 = static_cast<unsigned short>(*_Mid1);
|
|
bool _Save = false;
|
|
|
|
if (1u < *_Pstate) { // get saved MS 11 bits from *_Pstate
|
|
if (_Ch1 < 0xdc00u || 0xe000u <= _Ch1) {
|
|
return _Mybase::error; // bad second word
|
|
}
|
|
|
|
_Ch = static_cast<unsigned long>((*_Pstate << 10) | (_Ch1 - 0xdc00));
|
|
} else if (0xd800u <= _Ch1 && _Ch1 < 0xdc00u) { // get new first word
|
|
_Ch = static_cast<unsigned long>((_Ch1 - 0xd800 + 0x0040) << 10);
|
|
_Save = true; // put only first byte, rest with second word
|
|
} else {
|
|
_Ch = _Ch1; // not first word, just put it
|
|
}
|
|
|
|
_Byte _By;
|
|
int _Nextra;
|
|
|
|
if (_Ch < 0x0080u) {
|
|
_By = static_cast<_Byte>(_Ch);
|
|
_Nextra = 0;
|
|
} else if (_Ch < 0x0800u) {
|
|
_By = static_cast<_Byte>(0xc0 | _Ch >> 6);
|
|
_Nextra = 1;
|
|
} else if (_Ch < 0x10000u) {
|
|
_By = static_cast<_Byte>(0xe0 | _Ch >> 12);
|
|
_Nextra = 2;
|
|
} else {
|
|
_By = static_cast<_Byte>(0xf0 | _Ch >> 18);
|
|
_Nextra = 3;
|
|
}
|
|
|
|
int _Nput = _Nextra < 3 ? _Nextra + 1 : _Save ? 1 : 3;
|
|
|
|
if (_Last2 - _Mid2 < _Nput) {
|
|
break; // not enough room, even without header
|
|
}
|
|
|
|
if constexpr ((_Mymode & generate_header) != 0) { // header to put
|
|
if (*_Pstate == 0u) {
|
|
if (_Last2 - _Mid2 < 3 + _Nput) {
|
|
break; // not enough room for header + output
|
|
}
|
|
|
|
// prepend header
|
|
*_Mid2++ = '\xef';
|
|
*_Mid2++ = '\xbb';
|
|
*_Mid2++ = '\xbf';
|
|
}
|
|
}
|
|
|
|
++_Mid1;
|
|
if (_Save || _Nextra < 3) { // put first byte of sequence, if not already put
|
|
*_Mid2++ = _By;
|
|
--_Nput;
|
|
}
|
|
|
|
for (; 0 < _Nput; --_Nput) {
|
|
*_Mid2++ = static_cast<_Byte>((_Ch >> 6 * --_Nextra & 0x3f) | 0x80);
|
|
}
|
|
|
|
*_Pstate = static_cast<unsigned short>(_Save ? _Ch >> 10 : 1);
|
|
}
|
|
|
|
return _First1 == _Mid1 ? _Mybase::partial : _Mybase::ok;
|
|
}
|
|
|
|
result __CLR_OR_THIS_CALL do_unshift(_Statype&, _Byte* _First2, _Byte*, _Byte*& _Mid2) const override {
|
|
// generate bytes to return to default shift state
|
|
_Mid2 = _First2;
|
|
return _Mybase::noconv;
|
|
}
|
|
|
|
friend int _Codecvt_do_length<>(const codecvt_utf8_utf16&, _Statype&, const _Byte*, const _Byte*, size_t);
|
|
|
|
int __CLR_OR_THIS_CALL do_length(
|
|
_Statype& _State, const _Byte* _First1, const _Byte* _Last1, size_t _Count) const noexcept override {
|
|
return _Codecvt_do_length(*this, _State, _First1, _Last1, _Count);
|
|
}
|
|
|
|
bool __CLR_OR_THIS_CALL do_always_noconv() const noexcept override {
|
|
// return true if conversions never change input
|
|
return false;
|
|
}
|
|
|
|
int __CLR_OR_THIS_CALL do_max_length() const noexcept override {
|
|
// return maximum length required for a conversion
|
|
if constexpr ((_Mymode & consume_header) != 0) {
|
|
return 9; // header + max input
|
|
} else if constexpr ((_Mymode & generate_header) != 0) {
|
|
return 7; // header + max output
|
|
} else {
|
|
return 6; // 6-byte max input sequence, no 3-byte header
|
|
}
|
|
}
|
|
|
|
int __CLR_OR_THIS_CALL do_encoding() const noexcept override {
|
|
// return length of code sequence (from codecvt)
|
|
return 0; // 0 => varying length
|
|
}
|
|
};
|
|
_STL_RESTORE_DEPRECATED_WARNING
|
|
_STD_END
|
|
#pragma pop_macro("new")
|
|
_STL_RESTORE_CLANG_WARNINGS
|
|
#pragma warning(pop)
|
|
#pragma pack(pop)
|
|
#endif // _STL_COMPILER_PREPROCESSOR
|
|
#endif // _CODECVT_
|