blob: d8ba2b39c18559fe4661713288f665767417ce3d [file] [log] [blame]
/*
* Copyright 2006 The Android Open Source Project
*
* Use of this source code is governed by a BSD-style license that can be
* found in the LICENSE file.
*/
#include "SkUtils.h"
/* 0xxxxxxx 1 total
10xxxxxx // never a leading byte
110xxxxx 2 total
1110xxxx 3 total
11110xxx 4 total
11 10 01 01 xx xx xx xx 0...
0xE5XX0000
0xE5 << 24
*/
static bool utf8_byte_is_valid(uint8_t c) {
return c < 0xF5 && (c & 0xFE) != 0xC0;
}
static bool utf8_byte_is_continuation(uint8_t c) {
return (c & 0xC0) == 0x80;
}
static bool utf8_byte_is_leading_byte(uint8_t c) {
return utf8_byte_is_valid(c) && !utf8_byte_is_continuation(c);
}
#ifdef SK_DEBUG
static void assert_utf8_leadingbyte(unsigned c) {
SkASSERT(utf8_byte_is_leading_byte(SkToU8(c)));
}
int SkUTF8_LeadByteToCount(unsigned c) {
assert_utf8_leadingbyte(c);
return (((0xE5 << 24) >> (c >> 4 << 1)) & 3) + 1;
}
#else
#define assert_utf8_leadingbyte(c)
#endif
/**
* @returns -1 iff invalid UTF8 byte,
* 0 iff UTF8 continuation byte,
* 1 iff ASCII byte,
* 2 iff leading byte of 2-byte sequence,
* 3 iff leading byte of 3-byte sequence, and
* 4 iff leading byte of 4-byte sequence.
*
* I.e.: if return value > 0, then gives length of sequence.
*/
static int utf8_byte_type(uint8_t c) {
if (c < 0x80) {
return 1;
} else if (c < 0xC0) {
return 0;
} else if (c < 0xF5 && (c & 0xFE) != 0xC0) { // "octet values C0, C1, F5 to FF never appear"
return (((0xE5 << 24) >> ((unsigned)c >> 4 << 1)) & 3) + 1;
} else {
return -1;
}
}
static bool utf8_type_is_valid_leading_byte(int type) { return type > 0; }
int SkUTF8_CountUnichars(const char utf8[]) {
SkASSERT(utf8);
int count = 0;
for (;;) {
int c = *(const uint8_t*)utf8;
if (c == 0) {
break;
}
utf8 += SkUTF8_LeadByteToCount(c);
count += 1;
}
return count;
}
// SAFE: returns -1 if invalid UTF-8
int SkUTF8_CountUnichars(const void* text, size_t byteLength) {
SkASSERT(text);
const char* utf8 = static_cast<const char*>(text);
if (byteLength == 0) {
return 0;
}
int count = 0;
const char* stop = utf8 + byteLength;
while (utf8 < stop) {
int type = utf8_byte_type(*(const uint8_t*)utf8);
SkASSERT(type >= -1 && type <= 4);
if (!utf8_type_is_valid_leading_byte(type) || utf8 + type > stop) {
// Sequence extends beyond end.
return -1;
}
while(type-- > 1) {
++utf8;
if (!utf8_byte_is_continuation(*(const uint8_t*)utf8)) {
return -1;
}
}
++utf8;
++count;
}
return count;
}
SkUnichar SkUTF8_ToUnichar(const char utf8[]) {
SkASSERT(utf8);
const uint8_t* p = (const uint8_t*)utf8;
int c = *p;
int hic = c << 24;
assert_utf8_leadingbyte(c);
if (hic < 0) {
uint32_t mask = (uint32_t)~0x3F;
hic = SkLeftShift(hic, 1);
do {
c = (c << 6) | (*++p & 0x3F);
mask <<= 5;
} while ((hic = SkLeftShift(hic, 1)) < 0);
c &= ~mask;
}
return c;
}
// SAFE: returns -1 on invalid UTF-8 sequence.
SkUnichar SkUTF8_NextUnicharWithError(const char** ptr, const char* end) {
SkASSERT(ptr && *ptr);
SkASSERT(*ptr < end);
const uint8_t* p = (const uint8_t*)*ptr;
int c = *p;
int hic = c << 24;
if (!utf8_byte_is_leading_byte(c)) {
return -1;
}
if (hic < 0) {
uint32_t mask = (uint32_t)~0x3F;
hic = SkLeftShift(hic, 1);
do {
++p;
if (p >= (const uint8_t*)end) {
return -1;
}
// check before reading off end of array.
uint8_t nextByte = *p;
if (!utf8_byte_is_continuation(nextByte)) {
return -1;
}
c = (c << 6) | (nextByte & 0x3F);
mask <<= 5;
} while ((hic = SkLeftShift(hic, 1)) < 0);
c &= ~mask;
}
*ptr = (char*)p + 1;
return c;
}
SkUnichar SkUTF8_NextUnichar(const char** ptr) {
SkASSERT(ptr && *ptr);
const uint8_t* p = (const uint8_t*)*ptr;
int c = *p;
int hic = c << 24;
assert_utf8_leadingbyte(c);
if (hic < 0) {
uint32_t mask = (uint32_t)~0x3F;
hic = SkLeftShift(hic, 1);
do {
c = (c << 6) | (*++p & 0x3F);
mask <<= 5;
} while ((hic = SkLeftShift(hic, 1)) < 0);
c &= ~mask;
}
*ptr = (char*)p + 1;
return c;
}
SkUnichar SkUTF8_PrevUnichar(const char** ptr) {
SkASSERT(ptr && *ptr);
const char* p = *ptr;
if (*--p & 0x80) {
while (*--p & 0x40) {
;
}
}
*ptr = (char*)p;
return SkUTF8_NextUnichar(&p);
}
size_t SkUTF8_FromUnichar(SkUnichar uni, char utf8[]) {
if ((uint32_t)uni > 0x10FFFF) {
SkDEBUGFAIL("bad unichar");
return 0;
}
if (uni <= 127) {
if (utf8) {
*utf8 = (char)uni;
}
return 1;
}
char tmp[4];
char* p = tmp;
size_t count = 1;
SkDEBUGCODE(SkUnichar orig = uni;)
while (uni > 0x7F >> count) {
*p++ = (char)(0x80 | (uni & 0x3F));
uni >>= 6;
count += 1;
}
if (utf8) {
p = tmp;
utf8 += count;
while (p < tmp + count - 1) {
*--utf8 = *p++;
}
*--utf8 = (char)(~(0xFF >> count) | uni);
}
SkASSERT(utf8 == nullptr || orig == SkUTF8_ToUnichar(utf8));
return count;
}
///////////////////////////////////////////////////////////////////////////////
int SkUTF16_CountUnichars(const uint16_t src[]) {
SkASSERT(src);
int count = 0;
unsigned c;
while ((c = *src++) != 0) {
SkASSERT(!SkUTF16_IsLowSurrogate(c));
if (SkUTF16_IsHighSurrogate(c)) {
c = *src++;
SkASSERT(SkUTF16_IsLowSurrogate(c));
}
count += 1;
}
return count;
}
// returns -1 on error
int SkUTF16_CountUnichars(const void* text, size_t byteLength) {
SkASSERT(text);
if (byteLength == 0) {
return 0;
}
if (!SkIsAlign2(intptr_t(text)) || !SkIsAlign2(byteLength)) {
return -1;
}
const uint16_t* src = static_cast<const uint16_t*>(text);
const uint16_t* stop = src + (byteLength >> 1);
int count = 0;
while (src < stop) {
unsigned c = *src++;
SkASSERT(!SkUTF16_IsLowSurrogate(c));
if (SkUTF16_IsHighSurrogate(c)) {
if (src >= stop) {
return -1;
}
c = *src++;
if (!SkUTF16_IsLowSurrogate(c)) {
return -1;
}
}
count += 1;
}
return count;
}
SkUnichar SkUTF16_NextUnichar(const uint16_t** srcPtr) {
SkASSERT(srcPtr && *srcPtr);
const uint16_t* src = *srcPtr;
SkUnichar c = *src++;
SkASSERT(!SkUTF16_IsLowSurrogate(c));
if (SkUTF16_IsHighSurrogate(c)) {
unsigned c2 = *src++;
SkASSERT(SkUTF16_IsLowSurrogate(c2));
// c = ((c & 0x3FF) << 10) + (c2 & 0x3FF) + 0x10000
// c = (((c & 0x3FF) + 64) << 10) + (c2 & 0x3FF)
c = (c << 10) + c2 + (0x10000 - (0xD800 << 10) - 0xDC00);
}
*srcPtr = src;
return c;
}
SkUnichar SkUTF16_PrevUnichar(const uint16_t** srcPtr) {
SkASSERT(srcPtr && *srcPtr);
const uint16_t* src = *srcPtr;
SkUnichar c = *--src;
SkASSERT(!SkUTF16_IsHighSurrogate(c));
if (SkUTF16_IsLowSurrogate(c)) {
unsigned c2 = *--src;
SkASSERT(SkUTF16_IsHighSurrogate(c2));
c = (c2 << 10) + c + (0x10000 - (0xD800 << 10) - 0xDC00);
}
*srcPtr = src;
return c;
}
size_t SkUTF16_FromUnichar(SkUnichar uni, uint16_t dst[]) {
SkASSERT((unsigned)uni <= 0x10FFFF);
int extra = (uni > 0xFFFF);
if (dst) {
if (extra) {
// dst[0] = SkToU16(0xD800 | ((uni - 0x10000) >> 10));
// dst[0] = SkToU16(0xD800 | ((uni >> 10) - 64));
dst[0] = SkToU16((0xD800 - 64) + (uni >> 10));
dst[1] = SkToU16(0xDC00 | (uni & 0x3FF));
SkASSERT(SkUTF16_IsHighSurrogate(dst[0]));
SkASSERT(SkUTF16_IsLowSurrogate(dst[1]));
} else {
dst[0] = SkToU16(uni);
SkASSERT(!SkUTF16_IsHighSurrogate(dst[0]));
SkASSERT(!SkUTF16_IsLowSurrogate(dst[0]));
}
}
return 1 + extra;
}
size_t SkUTF16_ToUTF8(const uint16_t utf16[], int numberOf16BitValues,
char utf8[]) {
SkASSERT(numberOf16BitValues >= 0);
if (numberOf16BitValues <= 0) {
return 0;
}
SkASSERT(utf16 != nullptr);
const uint16_t* stop = utf16 + numberOf16BitValues;
size_t size = 0;
if (utf8 == nullptr) { // just count
while (utf16 < stop) {
size += SkUTF8_FromUnichar(SkUTF16_NextUnichar(&utf16), nullptr);
}
} else {
char* start = utf8;
while (utf16 < stop) {
utf8 += SkUTF8_FromUnichar(SkUTF16_NextUnichar(&utf16), utf8);
}
size = utf8 - start;
}
return size;
}
const char SkHexadecimalDigits::gUpper[16] =
{ '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };
const char SkHexadecimalDigits::gLower[16] =
{ '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f' };
// returns -1 on error
int SkUTF32_CountUnichars(const void* text, size_t byteLength) {
if (byteLength == 0) {
return 0;
}
if (!SkIsAlign4(intptr_t(text)) || !SkIsAlign4(byteLength)) {
return -1;
}
const uint32_t kInvalidUnicharMask = 0xFF000000; // unichar fits in 24 bits
const uint32_t* ptr = static_cast<const uint32_t*>(text);
const uint32_t* stop = ptr + (byteLength >> 2);
while (ptr < stop) {
if (*ptr & kInvalidUnicharMask) {
return -1;
}
ptr += 1;
}
return SkToInt(byteLength >> 2);
}