mirror of
https://github.com/CloverHackyColor/CloverBootloader.git
synced 2024-11-10 09:40:53 +01:00
402 lines
8.8 KiB
C++
402 lines
8.8 KiB
C++
|
//
|
||
|
// utf8Conversion.hpp
|
||
|
//
|
||
|
// Created by jief the 24 Feb 2020.
|
||
|
//
|
||
|
|
||
|
#include "unicode_conversions.h"
|
||
|
|
||
|
#ifndef MIN
|
||
|
#define MIN(a, b) ((a) < (b) ? (a) : (b))
|
||
|
#endif
|
||
|
|
||
|
|
||
|
|
||
|
size_t char16_len(const char16_t* s)
|
||
|
{
|
||
|
const char16_t* p = s;
|
||
|
while ( *p++ );
|
||
|
return (size_t)(p-s-1);
|
||
|
}
|
||
|
|
||
|
size_t char32_len(const char32_t* s)
|
||
|
{
|
||
|
const char32_t* p = s;
|
||
|
while ( *p++ );
|
||
|
return (size_t)(p-s-1);
|
||
|
}
|
||
|
|
||
|
size_t wchar_len(const wchar_t* s)
|
||
|
{
|
||
|
#if __WCHAR_MAX__ <= 0xFFFFu
|
||
|
return char16_len((const char16_t*)s);
|
||
|
#else
|
||
|
return char32_len((const char32_t*)s);
|
||
|
#endif
|
||
|
}
|
||
|
|
||
|
|
||
|
static inline int is_surrogate(char16_t uc) { return (uc - 0xd800u) < 2048u; }
|
||
|
static inline int is_high_surrogate(char16_t uc) { return (uc & 0xfffffc00) == 0xd800; }
|
||
|
static inline int is_low_surrogate(char16_t uc) { return (uc & 0xfffffc00) == 0xdc00; }
|
||
|
|
||
|
static inline char32_t surrogate_to_utf32(char16_t high, char16_t low) {
|
||
|
return char32_t((high << 10) + low - 0x35fdc00); // Safe cast, it fits in 32 bits
|
||
|
}
|
||
|
|
||
|
|
||
|
/************************************************************* Char conversion *********************************************************/
|
||
|
|
||
|
#define halfBase 0x0010000UL
|
||
|
#define halfMask 0x3FFUL
|
||
|
#define halfShift 10 /* used for shifting by 10 bits */
|
||
|
#define UNI_SUR_HIGH_START 0xD800u
|
||
|
#define UNI_SUR_LOW_START 0xDC00u
|
||
|
|
||
|
|
||
|
|
||
|
const char* get_char32_from_utf8_string(const char *s, char32_t* char32)
|
||
|
{
|
||
|
char32_t c;
|
||
|
if (*s & 0x80) {
|
||
|
if (*(s+1) == 0) {
|
||
|
// Finished in the middle of an utf8 multibyte char
|
||
|
return 0;
|
||
|
}
|
||
|
if ((*(s+1) & 0xc0) != 0x80) {
|
||
|
s += 1;
|
||
|
return 0;
|
||
|
}
|
||
|
if ((*s & 0xe0) == 0xe0) {
|
||
|
if (*(s+2) == 0) {
|
||
|
// Finished in the middle of an utf8 multibyte char
|
||
|
return 0;
|
||
|
}
|
||
|
if ((*(s+2) & 0xc0) != 0x80) {
|
||
|
s += 2;
|
||
|
return 0;
|
||
|
}
|
||
|
if ((*s & 0xf0) == 0xf0) {
|
||
|
if (*(s+3) == 0) {
|
||
|
// Finished in the middle of an utf8 multibyte char
|
||
|
return 0;
|
||
|
}
|
||
|
if ((*s & 0xf8) != 0xf0 || (*(s+3) & 0xc0) != 0x80) {
|
||
|
s += 3;
|
||
|
return 0;
|
||
|
}
|
||
|
/* 4-byte code */
|
||
|
c = char32_t((*s & 0x7) << 18); // & result type is int. We know it fits in 32 bits. Safe to cast to char32_t
|
||
|
c |= char32_t((*(s+1) & 0x3f) << 12);
|
||
|
c |= char32_t((*(s+2) & 0x3f) << 6);
|
||
|
c |= *(s+3) & 0x3f;
|
||
|
s += 4;
|
||
|
} else {
|
||
|
/* 3-byte code */
|
||
|
c = char32_t((*s & 0xf) << 12);
|
||
|
c |= char32_t((*(s+1) & 0x3f) << 6);
|
||
|
c |= *(s+2) & 0x3f;
|
||
|
s += 3;
|
||
|
}
|
||
|
} else {
|
||
|
/* 2-byte code */
|
||
|
c = char32_t((*s & 0x1f) << 6);
|
||
|
c |= *(s+1) & 0x3f;
|
||
|
s += 2;
|
||
|
}
|
||
|
} else {
|
||
|
/* 1-byte code */
|
||
|
c = (unsigned char)(*s); // in case we compiled with signed char
|
||
|
s += 1;
|
||
|
}
|
||
|
*char32 = c;
|
||
|
return s;
|
||
|
}
|
||
|
|
||
|
void get_char16_from_char32(char32_t char32, char16_t* char16_1, char16_t* char16_2)
|
||
|
{
|
||
|
if ( char32 <= 0xFFFF) {
|
||
|
*char16_1 = (char16_t)char32;
|
||
|
*char16_2 = 0;
|
||
|
} else {
|
||
|
char32 -= halfBase;
|
||
|
*char16_1 = (char16_t)((char32 >> halfShift) + UNI_SUR_HIGH_START);
|
||
|
*char16_2 = (char16_t)((char32 & halfMask) + UNI_SUR_LOW_START);
|
||
|
}
|
||
|
}
|
||
|
|
||
|
char32_t get_char32_from_char16(char16_t char16_1, char16_t char16_2)
|
||
|
{
|
||
|
if (!is_surrogate(char16_1)) {
|
||
|
return char16_1;
|
||
|
} else {
|
||
|
if (is_high_surrogate(char16_1) && is_low_surrogate(char16_2)) {
|
||
|
return surrogate_to_utf32(char16_1, char16_2);
|
||
|
} else {
|
||
|
return 0;
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
|
||
|
const char16_t* get_char32_from_char16_string(const char16_t* s, char32_t* char32)
|
||
|
{
|
||
|
const char16_t char16_1 = *s++;
|
||
|
if (!is_surrogate(char16_1)) {
|
||
|
*char32 = char16_1;
|
||
|
return s;
|
||
|
} else {
|
||
|
if (is_high_surrogate(char16_1) && is_low_surrogate(*s)) {
|
||
|
*char32 = surrogate_to_utf32(char16_1, *s++);
|
||
|
return s;
|
||
|
} else {
|
||
|
return 0;
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
|
||
|
/*
|
||
|
* dst_max_len MUST be >= 1 when called
|
||
|
*/
|
||
|
char* get_utf8_from_char32(char* dst, size_t* dst_max_len, char32_t utf32_char)
|
||
|
{
|
||
|
#ifdef JIEF_DEBUG
|
||
|
char* dst_debug = dst;
|
||
|
(void)dst_debug;
|
||
|
#endif
|
||
|
/* assertion: utf32_char is a single UTF-4 value */
|
||
|
/* assertion: dst_max_len >= 1 */
|
||
|
int bits = 0; // just to silence the warning
|
||
|
|
||
|
if (utf32_char < 0x80) {
|
||
|
*dst++ = (char)utf32_char;
|
||
|
*dst_max_len -= 1;
|
||
|
bits = -6;
|
||
|
}
|
||
|
else if (utf32_char < 0x800) {
|
||
|
*dst++ = (char)(((utf32_char >> 6) & 0x1F) | 0xC0);
|
||
|
*dst_max_len -= 1;
|
||
|
bits = 0;
|
||
|
}
|
||
|
else if (utf32_char < 0x10000) {
|
||
|
*dst++ = (char)(((utf32_char >> 12) & 0x0F) | 0xE0);
|
||
|
*dst_max_len -= 1;
|
||
|
bits = 6;
|
||
|
}
|
||
|
else {
|
||
|
*dst++ = (char)(((utf32_char >> 18) & 0x07) | 0xF0);
|
||
|
*dst_max_len -= 1;
|
||
|
bits = 12;
|
||
|
}
|
||
|
for ( ; *dst_max_len > 0 && bits >= 0 ; bits -= 6 ) {
|
||
|
*dst++ = (char)(((utf32_char >> bits) & 0x3F) | 0x80);
|
||
|
*dst_max_len -= 1;
|
||
|
}
|
||
|
#ifdef JIEF_DEBUG
|
||
|
*dst = 0;
|
||
|
#endif
|
||
|
return dst;
|
||
|
}
|
||
|
|
||
|
|
||
|
/************************************************************* utf8 - char16 *********************************************************/
|
||
|
|
||
|
|
||
|
size_t utf8_string_char16_count(const char *s)
|
||
|
{
|
||
|
if ( !s ) return 0;
|
||
|
size_t len = 0;
|
||
|
|
||
|
while ( *s ) {
|
||
|
char32_t c;
|
||
|
s = get_char32_from_utf8_string(s, &c);
|
||
|
if ( c == 0 ) return len;
|
||
|
if ( c <= 0xFFFF) {
|
||
|
len += 1;
|
||
|
}else{
|
||
|
len += 2;
|
||
|
}
|
||
|
}
|
||
|
return len;
|
||
|
}
|
||
|
|
||
|
size_t char16_string_from_utf8_string(char16_t* dst, size_t dst_max_len, const char *s)
|
||
|
{
|
||
|
if ( dst_max_len == 0 ) return 0;
|
||
|
dst_max_len -= 1;
|
||
|
|
||
|
// size_t dst_len = 0;
|
||
|
char16_t* p = dst;
|
||
|
char16_t* p_max = dst + dst_max_len;
|
||
|
|
||
|
while ( *s && p < p_max ) {
|
||
|
char32_t c;
|
||
|
s = get_char32_from_utf8_string(s, &c);
|
||
|
if ( c == 0 ) return (size_t)(p-dst);
|
||
|
char16_t char16_1, char16_2;
|
||
|
get_char16_from_char32(c, &char16_1, &char16_2);
|
||
|
if ( char16_2 != 0 ) {
|
||
|
if ( p < p_max-1 ) {
|
||
|
*p++ = char16_1;
|
||
|
*p++ = char16_2;
|
||
|
}else{
|
||
|
*p = 0;
|
||
|
return (size_t)(p-dst);
|
||
|
}
|
||
|
}else{
|
||
|
*p++ = char16_1;
|
||
|
}
|
||
|
}
|
||
|
*p = 0;
|
||
|
return (size_t)(p-dst);
|
||
|
}
|
||
|
|
||
|
size_t utf8_string_from_char16_string(char* dst, size_t dst_max_len, const char16_t *s)
|
||
|
{
|
||
|
char* p = dst;
|
||
|
while ( *s && dst_max_len > 0 ) {
|
||
|
char32_t utf32_char;
|
||
|
s = get_char32_from_char16_string(s, &utf32_char);
|
||
|
p = get_utf8_from_char32(p, &dst_max_len, utf32_char);
|
||
|
}
|
||
|
*p = 0;
|
||
|
return (size_t)(p-dst);
|
||
|
}
|
||
|
|
||
|
|
||
|
|
||
|
/************************************************************* utf8 - char32 *********************************************************/
|
||
|
|
||
|
|
||
|
size_t utf8_string_char32_count(const char *s)
|
||
|
{
|
||
|
if ( !s ) return 0;
|
||
|
size_t len = 0;
|
||
|
|
||
|
while ( *s ) {
|
||
|
char32_t c;
|
||
|
s = get_char32_from_utf8_string(s, &c);
|
||
|
if ( c == 0 ) return len;
|
||
|
len += 1;
|
||
|
}
|
||
|
return len;
|
||
|
}
|
||
|
|
||
|
|
||
|
size_t char32_string_from_utf8_string(char32_t* dst, size_t dst_max_len, const char *s)
|
||
|
{
|
||
|
if ( dst_max_len == 0 ) return 0;
|
||
|
char32_t* p = dst;
|
||
|
char32_t* p_max = dst + dst_max_len - 1;
|
||
|
|
||
|
while ( *s && p < p_max ) {
|
||
|
char32_t c;
|
||
|
s = get_char32_from_utf8_string(s, &c);
|
||
|
if ( c == 0 ) return (size_t)(p-dst);
|
||
|
*p++ = c;
|
||
|
}
|
||
|
*p = 0;
|
||
|
return (size_t)(p-dst);
|
||
|
}
|
||
|
|
||
|
size_t utf8_string_from_char32_string(char* dst, size_t dst_max_len, const char32_t *s)
|
||
|
{
|
||
|
char* p = dst;
|
||
|
while ( *s && dst_max_len > 0 ) {
|
||
|
p = get_utf8_from_char32(p, &dst_max_len, *s++);
|
||
|
}
|
||
|
*p = 0;
|
||
|
return (size_t)(p-dst);
|
||
|
}
|
||
|
|
||
|
|
||
|
/************************************************************* utf8 - wchar *********************************************************/
|
||
|
|
||
|
size_t utf8_string_wchar_count(const char *s)
|
||
|
{
|
||
|
#if __WCHAR_MAX__ <= 0xFFFFu
|
||
|
return utf8_string_char16_count(s);
|
||
|
#else
|
||
|
return utf8_string_char32_count(s);
|
||
|
#endif
|
||
|
}
|
||
|
|
||
|
size_t wchar_string_from_utf8_string(wchar_t* dst, size_t dst_max_len, const char *s)
|
||
|
{
|
||
|
#if __WCHAR_MAX__ <= 0xFFFFu
|
||
|
return char16_string_from_utf8_string((char16_t*)dst, dst_max_len, s);
|
||
|
#else
|
||
|
return char32_string_from_utf8_string((char32_t*)dst, dst_max_len, s);
|
||
|
#endif
|
||
|
}
|
||
|
|
||
|
size_t utf8_string_from_wchar_string(char* dst, size_t dst_max_len, const wchar_t* s)
|
||
|
{
|
||
|
#if __WCHAR_MAX__ <= 0xFFFFu
|
||
|
return utf8_string_from_char16_string(dst, dst_max_len, (char16_t*)s);
|
||
|
#else
|
||
|
return utf8_string_from_char32_string(dst, dst_max_len, (char32_t*)s);
|
||
|
#endif
|
||
|
}
|
||
|
|
||
|
|
||
|
|
||
|
|
||
|
/************************************************************* char16 - char32 *********************************************************/
|
||
|
|
||
|
|
||
|
size_t char16_string_char32_count(const char16_t *s)
|
||
|
{
|
||
|
if ( !s ) return 0;
|
||
|
size_t len = 0;
|
||
|
|
||
|
while ( *s ) {
|
||
|
char32_t c;
|
||
|
s = get_char32_from_char16_string(s, &c);
|
||
|
if ( c == 0 ) return len;
|
||
|
len += 1;
|
||
|
}
|
||
|
return len;
|
||
|
}
|
||
|
|
||
|
|
||
|
size_t utf32_string_to_char16_string(char32_t* dst, size_t dst_max_len, const char16_t *s)
|
||
|
{
|
||
|
if ( dst_max_len == 0 ) return 0;
|
||
|
char32_t* p = dst;
|
||
|
char32_t* p_max = dst + dst_max_len - 1;
|
||
|
|
||
|
while ( *s && p < p_max ) {
|
||
|
char32_t c;
|
||
|
s = get_char32_from_char16_string(s, &c);
|
||
|
if ( c == 0 ) return (size_t)(p-dst);
|
||
|
*p++ = c;
|
||
|
}
|
||
|
*p = 0;
|
||
|
return (size_t)(p-dst);
|
||
|
}
|
||
|
|
||
|
size_t utf16_string_to_char32_string(char16_t* dst, size_t dst_max_len, const char32_t *s)
|
||
|
{
|
||
|
if ( dst_max_len == 0 ) return 0;
|
||
|
char16_t* p = dst;
|
||
|
char16_t* p_max = dst + dst_max_len - 1;
|
||
|
|
||
|
while ( *s && p < p_max ) {
|
||
|
char16_t char16_1, char16_2;
|
||
|
get_char16_from_char32(*s++, &char16_1, &char16_2);
|
||
|
if ( char16_2 != 0 ) {
|
||
|
if ( p < p_max-1 ) {
|
||
|
*p++ = char16_1;
|
||
|
*p++ = char16_2;
|
||
|
}else{
|
||
|
*p = 0;
|
||
|
return (size_t)(p-dst);
|
||
|
}
|
||
|
}else{
|
||
|
*p++ = char16_1;
|
||
|
}
|
||
|
}
|
||
|
*p = 0;
|
||
|
return (size_t)(p-dst);
|
||
|
}
|