scummvm/common/str-enc.cpp
2021-07-22 00:06:25 +03:00

932 lines
21 KiB
C++

/* ScummVM - Graphic Adventure Engine
*
* ScummVM is the legal property of its developers, whose names
* are too numerous to list here. Please refer to the COPYRIGHT
* file distributed with this source distribution.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
*/
#include "common/str.h"
#include "common/ustr.h"
#include "common/util.h"
#include "common/endian.h"
#include "common/error.h"
#include "common/system.h"
#include "common/enc-internal.h"
#include "common/file.h"
namespace Common {
// //TODO: This is a quick and dirty converter. Refactoring needed:
// 1. Original version has an option for performing strict / nonstrict
// conversion for the 0xD800...0xDFFF interval
// 2. Original version returns a result code. This version does NOT
// insert 'FFFD' on errors & does not inform caller on any errors
//
// More comprehensive one lives in wintermute/utils/convert_utf.cpp
void U32String::decodeUTF8(const char *src, uint32 len) {
ensureCapacity(len, false);
// The String class, and therefore the Font class as well, assume one
// character is one byte, but in this case it's actually an UTF-8
// string with up to 4 bytes per character. To work around this,
// convert it to an U32String before drawing it, because our Font class
// can handle that.
for (uint i = 0; i < len;) {
uint32 chr = 0;
uint num = 1;
if ((src[i] & 0xF8) == 0xF0) {
num = 4;
} else if ((src[i] & 0xF0) == 0xE0) {
num = 3;
} else if ((src[i] & 0xE0) == 0xC0) {
num = 2;
}
if (len - i >= num) {
switch (num) {
case 4:
chr |= (src[i++] & 0x07) << 18;
chr |= (src[i++] & 0x3F) << 12;
chr |= (src[i++] & 0x3F) << 6;
chr |= (src[i++] & 0x3F);
break;
case 3:
chr |= (src[i++] & 0x0F) << 12;
chr |= (src[i++] & 0x3F) << 6;
chr |= (src[i++] & 0x3F);
break;
case 2:
chr |= (src[i++] & 0x1F) << 6;
chr |= (src[i++] & 0x3F);
break;
default:
chr = (src[i++] & 0x7F);
break;
}
} else {
break;
}
operator+=(chr);
}
}
const uint16 invalidCode = 0xFFFD;
static bool cjk_tables_loaded = false;
static const uint16 *windows932ConversionTable;
static const uint16 *windows949ConversionTable;
static const uint16 *windows950ConversionTable;
static const uint16 *loadCJKTable(File &f, int idx, size_t sz) {
f.seek(16 + idx * 4);
uint32 off = f.readUint32LE();
f.seek(off);
uint16 *res = new uint16[sz];
f.read(res, 2 * sz);
#ifndef SCUMM_LITTLE_ENDIAN
for (uint i = 0; i < sz; i++)
res[i] = FROM_LE_16(res[i]);
#endif
return res;
}
static void loadCJKTables() {
File f;
cjk_tables_loaded = true;
if (!f.open("encoding.dat")) {
warning("encoding.dat is not found. Support for CJK is disabled");
return;
}
if (f.size() < 16 + 3 * 4) {
warning("encoding.dat is invalid. Support for CJK is disabled");
return;
}
if (f.readUint32BE() != MKTAG('S', 'C', 'V', 'M')
|| f.readUint32BE() != MKTAG('E', 'N', 'C', 'D')) {
warning("encoding.dat is invalid. Support for CJK is disabled");
return;
}
// Version and number of tables.
if (f.readUint32LE() != 0 || f.readUint32LE() < 3) {
warning("encoding.dat is of incompatible version. Support for CJK is disabled");
return;
}
windows932ConversionTable = loadCJKTable(f, 0, 47 * 192);
windows949ConversionTable = loadCJKTable(f, 1, 0x7e * 0xb2);
windows950ConversionTable = loadCJKTable(f, 2, 89 * 157);
}
void U32String::decodeWindows932(const char *src, uint32 len) {
ensureCapacity(len, false);
if (!cjk_tables_loaded)
loadCJKTables();
for (uint i = 0; i < len;) {
uint8 high = src[i++];
if ((high & 0x80) == 0x00) {
operator+=(high);
continue;
}
// Katakana
if (high >= 0xa1 && high <= 0xdf) {
operator+=(high - 0xa1 + 0xFF61);
continue;
}
if (i >= len) {
operator+=(invalidCode);
continue;
}
uint8 low = src[i++];
if (low < 0x40) {
operator+=(invalidCode);
continue;
}
uint8 lowidx = low - 0x40;
uint8 highidx;
if (high >= 0x81 && high < 0x85)
highidx = high - 0x81;
else if (high >= 0x87 && high < 0xa0)
highidx = high - 0x87 + 4;
else if (high >= 0xe0 && high < 0xef)
highidx = high - 0xe0 + 29;
else if (high >= 0xfa && high < 0xfd)
highidx = high - 0xfa + 44;
else {
operator+=(invalidCode);
continue;
}
if (!windows932ConversionTable) {
operator+=(invalidCode);
continue;
}
// Main range
uint16 val = windows932ConversionTable[highidx * 192 + lowidx];
operator+=(val ? val : invalidCode);
}
}
static uint16 convertUHCToUCSReal(uint8 high, uint8 low) {
uint lowidx = 0;
if (low >= 0x41 && low < 0x5b)
lowidx = low - 0x41;
else if (low >= 0x61 && low < 0x7b)
lowidx = low - 0x61 + 0x1a;
else if (low >= 0x81 && low < 0xff)
lowidx = low - 0x81 + 0x1a * 2;
else
return 0;
if (!windows949ConversionTable)
return 0;
uint16 idx = (high - 0x81) * 0xb2 + lowidx;
return windows949ConversionTable[idx];
}
uint16 convertUHCToUCS(uint8 high, uint8 low) {
if (!cjk_tables_loaded)
loadCJKTables();
return convertUHCToUCSReal(high, low);
}
void U32String::decodeWindows949(const char *src, uint32 len) {
ensureCapacity(len, false);
if (!cjk_tables_loaded)
loadCJKTables();
for (uint i = 0; i < len;) {
uint8 high = src[i++];
if ((high & 0x80) == 0x00) {
operator+=(high);
continue;
}
if (high == 0x80 || high == 0xff) {
operator+=(invalidCode);
continue;
}
if (i >= len) {
operator+=(invalidCode);
continue;
}
uint8 low = src[i++];
uint16 val = convertUHCToUCSReal(high, low);
operator+=(val ? val : invalidCode);
}
}
void U32String::decodeWindows950(const char *src, uint32 len) {
ensureCapacity(len, false);
if (!cjk_tables_loaded)
loadCJKTables();
for (uint i = 0; i < len;) {
uint8 high = src[i++];
if ((high & 0x80) == 0x00) {
operator+=(high);
continue;
}
// Euro symbol
if (high == 0x80) {
operator+=(0x20ac);
continue;
}
if (high == 0xff) {
operator+=(invalidCode);
continue;
}
if (i >= len) {
operator+=(invalidCode);
continue;
}
uint8 low = src[i++];
uint8 lowidx = low < 0x80 ? low - 0x40 : (low - 0xa1 + 0x3f);
// Main range
if (high >= 0xa1 && high < 0xfa) {
uint16 val = windows950ConversionTable ?
windows950ConversionTable[(high - 0xa1) * 157 + lowidx] : 0;
operator+=(val ? val : invalidCode);
continue;
}
// PUA range
if (high <= 0x8d) {
operator+=(0xeeb8 + 157 * (high-0x81) + lowidx);
continue;
}
if (high <= 0xa0) {
operator+=(0xe311 + (157 * (high-0x8e)) + lowidx);
continue;
}
if (high >= 0xfa) {
operator+=(0xe000 + (157 * (high-0xfa)) + lowidx);
continue;
}
}
}
void String::encodeWindows932(const U32String &src) {
static uint16 *reverseTable;
ensureCapacity(src.size() * 2, false);
if (!cjk_tables_loaded)
loadCJKTables();
if (!reverseTable && windows932ConversionTable) {
uint16 *rt = new uint16[0x10000];
memset(rt, 0, sizeof(rt[0]) * 0x10000);
for (uint highidx = 0; highidx < 47; highidx++) {
uint8 high = 0;
if (highidx < 4)
high = highidx + 0x81;
else if (highidx < 29)
high = highidx + 0x87 - 4;
else if (highidx < 44)
high = highidx + 0xe0 - 29;
else
high = highidx + 0xfa - 44;
for (uint lowidx = 0; lowidx < 192; lowidx++) {
uint8 low = lowidx + 0x40;
uint16 unicode = windows932ConversionTable[highidx * 192 + lowidx];
rt[unicode] = (high << 8) | low;
}
}
reverseTable = rt;
}
for (uint i = 0; i < src.size();) {
uint32 point = src[i++];
if (point < 0x80) {
operator+=(point);
continue;
}
// Katakana
if (point >= 0xff61 && point <= 0xff9f) {
operator+=(0xa1 + (point - 0xFF61));
continue;
}
if (point > 0x10000) {
operator+=('?');
continue;
}
if (!reverseTable) {
operator+=('?');
continue;
}
uint16 rev = reverseTable[point];
if (rev != 0) {
operator+=(rev >> 8);
operator+=(rev & 0xff);
continue;
}
// This codepage contains cyrillic, so no need to transliterate
operator+=('?');
continue;
}
}
void String::encodeWindows949(const U32String &src) {
static const uint16 *reverseTable;
ensureCapacity(src.size() * 2, false);
if (!cjk_tables_loaded)
loadCJKTables();
if (!reverseTable && windows949ConversionTable) {
uint16 *rt = new uint16[0x10000];
memset(rt, 0, sizeof(rt[0]) * 0x10000);
for (uint lowidx = 0; lowidx < 0xb2; lowidx++) {
uint8 low = 0;
if (lowidx < 0x1a)
low = 0x41 + lowidx;
else if (lowidx < 0x1a * 2)
low = 0x61 + lowidx - 0x1a;
else
low = 0x81 + lowidx - 0x1a * 2;
for (uint highidx = 0; highidx < 0x7e; highidx++) {
uint8 high = highidx + 0x81;
uint16 unicode = windows949ConversionTable[highidx * 0xb2 + lowidx];
rt[unicode] = (high << 8) | low;
}
}
reverseTable = rt;
}
for (uint i = 0; i < src.size();) {
uint32 point = src[i++];
if (point < 0x80) {
operator+=(point);
continue;
}
if (point > 0x10000 || !reverseTable) {
operator+=('?');
continue;
}
uint16 rev = reverseTable[point];
if (rev == 0) {
// This codepage contains cyrillic, so no need to transliterate
operator+=('?');
continue;
}
operator+=(rev >> 8);
operator+=(rev & 0xff);
}
}
static const char g_cyrillicTransliterationTable[] = {
' ', 'E', 'D', 'G', 'E', 'Z', 'I', 'I', 'J', 'L', 'N', 'C', 'K', 'I', 'U', 'D',
'A', 'B', 'V', 'G', 'D', 'E', 'Z', 'Z', 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P',
'R', 'S', 'T', 'U', 'F', 'H', 'C', 'C', 'S', 'S', '\"', 'Y', '\'', 'E', 'U', 'A',
'a', 'b', 'v', 'g', 'd', 'e', 'z', 'z', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p',
'r', 's', 't', 'u', 'f', 'h', 'c', 'c', 's', 's', '\"', 'y', '\'', 'e', 'u', 'a',
'e', 'e', 'd', 'g', 'e', 'z', 'i', 'i', 'j', 'l', 'n', 'c', 'k', 'i', 'u', 'd',
};
void String::translitChar(U32String::value_type point) {
if (point == 0xa0) {
operator+=(' ');
return;
}
if (point == 0xad) {
operator+=('-');
return;
}
if (point == 0x2116) {
operator+=('N');
return;
}
if (point >= 0x401 && point <= 0x45f) {
operator+=(g_cyrillicTransliterationTable[point - 0x400]);
return;
}
operator+=('?');
}
void String::encodeWindows950(const U32String &src, bool transliterate) {
static uint16 *reverseTable;
ensureCapacity(src.size() * 2, false);
if (!cjk_tables_loaded)
loadCJKTables();
if (!reverseTable && windows950ConversionTable) {
uint16 *rt = new uint16[0x10000];
memset(rt, 0, sizeof(rt[0]) * 0x10000);
for (uint lowidx = 0; lowidx < 157; lowidx++) {
uint8 low = 0;
if (lowidx < 0x3f)
low = 0x40 + lowidx;
else
low = 0xa1 + lowidx - 0x3f;
for (uint highidx = 0; highidx < 89; highidx++) {
uint8 high = highidx + 0xa1;
uint16 unicode = windows950ConversionTable[highidx * 157 + lowidx];
rt[unicode] = (high << 8) | low;
}
}
reverseTable = rt;
}
for (uint i = 0; i < src.size();) {
uint32 point = src[i++];
if (point < 0x80) {
operator+=(point);
continue;
}
if (point > 0x10000) {
operator+=('?');
continue;
}
// Euro symbol
if (point == 0x20ac) {
operator+=((char) 0x80);
continue;
}
if (!reverseTable) {
operator+=('?');
continue;
}
uint16 rev = reverseTable[point];
if (rev != 0) {
operator+=(rev >> 8);
operator+=(rev & 0xff);
continue;
}
// PUA range
if (point >= 0xe000 && point <= 0xf848) {
byte lowidx = 0, high = 0, low = 0;
if (point <= 0xe310) {
high = (point - 0xe000) / 157 + 0xfa;
lowidx = (point - 0xe000) % 157;
} else if (point <= 0xeeb7) {
high = (point - 0xe311) / 157 + 0x8e;
lowidx = (point - 0xe311) % 157;
} else if (point <= 0xf6b0) {
high = (point - 0xeeb8) / 157 + 0x81;
lowidx = (point - 0xeeb8) % 157;
} else {
high = (point - 0xf672) / 157 + 0xc6;
lowidx = (point - 0xf672) % 157;
}
if (lowidx <= 0x3e)
low = 0x40 + lowidx;
else
low = 0x62 + lowidx;
operator+=(high);
operator+=(low);
reverseTable[point] = (high << 8) | low;
continue;
}
if (transliterate) {
translitChar(point);
continue;
}
operator+=('?');
continue;
}
}
// //TODO: This is a quick and dirty converter. Refactoring needed:
// 1. Original version has an option for performing strict / nonstrict
// conversion for the 0xD800...0xDFFF interval
// 2. Original version returns a result code. This version inserts '0xFFFD' if
// character does not fit in 4 bytes & does not inform caller on any errors
//
// More comprehensive one lives in wintermute/utils/convert_utf.cpp
void String::encodeUTF8(const U32String &src) {
ensureCapacity(src.size(), false);
static const uint8 firstByteMark[5] = { 0x00, 0x00, 0xC0, 0xE0, 0xF0 };
char writingBytes[5] = {0x00, 0x00, 0x00, 0x00, 0x00};
uint i = 0;
while (i < src.size()) {
unsigned short bytesToWrite = 0;
const uint32 byteMask = 0xBF;
const uint32 byteMark = 0x80;
uint32 ch = src[i++];
if (ch < (uint32)0x80) {
bytesToWrite = 1;
} else if (ch < (uint32)0x800) {
bytesToWrite = 2;
} else if (ch < (uint32)0x10000) {
bytesToWrite = 3;
} else if (ch <= 0x0010FFFF) {
bytesToWrite = 4;
} else {
bytesToWrite = 3;
ch = invalidCode;
}
char *pBytes = writingBytes + (4 - bytesToWrite);
switch (bytesToWrite) {
case 4:
pBytes[3] = (char)((ch | byteMark) & byteMask);
ch >>= 6;
// fallthrough
case 3:
pBytes[2] = (char)((ch | byteMark) & byteMask);
ch >>= 6;
// fallthrough
case 2:
pBytes[1] = (char)((ch | byteMark) & byteMask);
ch >>= 6;
// fallthrough
case 1:
pBytes[0] = (char)(ch | firstByteMark[bytesToWrite]);
break;
default:
break;
}
operator+=(pBytes);
}
}
#define decodeUTF16Template(suffix, read) \
Common::U32String U32String::decodeUTF16 ## suffix (const uint16 *start, uint len) { \
const uint16 *ptr = start; \
Common::U32String dst; \
dst.ensureCapacity(len, false); \
\
while (len > 0) { \
uint16 c = read(ptr++); \
len--; \
if (c >= 0xD800 && c <= 0xDBFF && len > 0) { \
uint16 low = read(ptr); \
if (low >= 0xDC00 && low <= 0xDFFF) { \
/* low is OK, we can advance pointer */ \
ptr++; len--; \
dst += ((c & 0x3ff) << 10) \
| (low & 0x3ff); \
} else { \
dst += invalidCode; \
} \
continue; \
} \
\
if (c >= 0xD800 && c <= 0xDFFF) { \
dst += invalidCode; \
continue; \
} \
dst += c; \
} \
\
return dst; \
}
decodeUTF16Template(BE, READ_BE_UINT16)
decodeUTF16Template(LE, READ_LE_UINT16)
decodeUTF16Template(Native, READ_UINT16)
#define encodeUTF16Template(suffix, write) \
uint16 *U32String::encodeUTF16 ## suffix (uint *len) const { \
uint16 *out = new uint16[_size * 2 + 1]; \
uint16 *ptr = out; \
\
for (uint i = 0; i < _size; i++) { \
uint32 c = _str[i]; \
if (c < 0x10000) { \
write(ptr++, c); \
continue; \
} \
write (ptr++, 0xD800 | ((c >> 10) & 0x3ff)); \
write (ptr++, 0xDC00 | (c & 0x3ff)); \
} \
\
write(ptr, 0); \
if (len) \
*len = ptr - out; \
\
return out; \
}
encodeUTF16Template(BE, WRITE_BE_UINT16)
encodeUTF16Template(LE, WRITE_LE_UINT16)
encodeUTF16Template(Native, WRITE_UINT16)
// Upper bound on unicode codepoint in any single-byte encoding. Must be divisible by 0x100 and be strictly above large codepoint
static const int kMaxCharSingleByte = 0x3000;
static const uint16 *
getConversionTable(CodePage page) {
switch (page) {
case kWindows1250:
return kWindows1250ConversionTable;
case kWindows1251:
return kWindows1251ConversionTable;
case kWindows1252:
return kWindows1252ConversionTable;
case kWindows1253:
return kWindows1253ConversionTable;
case kWindows1254:
return kWindows1254ConversionTable;
case kWindows1255:
return kWindows1255ConversionTable;
case kWindows1256:
return kWindows1256ConversionTable;
case kWindows1257:
return kWindows1257ConversionTable;
case kMacCentralEurope:
return kMacCentralEuropeConversionTable;
case kMacRoman:
return kMacRomanConversionTable;
case kISO8859_1:
return kLatin1ConversionTable;
case kISO8859_2:
return kLatin2ConversionTable;
case kISO8859_5:
return kISO5ConversionTable;
case kDos850:
return kDos850ConversionTable;
case kDos862:
return kDos862ConversionTable;
case kDos866:
return kDos866ConversionTable;
case kASCII:
return kASCIIConversionTable;
case kCodePageInvalid:
// Multibyte encodings. Can't be represented in simple table way
case kUtf8:
case kWindows932:
case kWindows949:
case kWindows950:
return nullptr;
}
return nullptr;
}
struct ReverseTablePrefixTreeLevel1 {
struct ReverseTablePrefixTreeLevel2 *next[kMaxCharSingleByte / 0x100];
bool valid;
};
struct ReverseTablePrefixTreeLevel2 {
uint8 end[256];
ReverseTablePrefixTreeLevel2() {
memset(end, 0, sizeof(end));
}
};
ReverseTablePrefixTreeLevel1 reverseTables[kLastEncoding + 1];
static const ReverseTablePrefixTreeLevel1 *
getReverseConversionTable(CodePage page) {
if (reverseTables[page].valid)
return &reverseTables[page];
const uint16 *conversionTable = getConversionTable(page);
if (!conversionTable)
return nullptr;
reverseTables[page].valid = true;
for (uint i = 0; i < 0x80; i++) {
uint32 c = conversionTable[i];
if (c == 0 || c >= kMaxCharSingleByte)
continue;
if (!reverseTables[page].next[c >> 8]) {
reverseTables[page].next[c >> 8] = new ReverseTablePrefixTreeLevel2();
}
reverseTables[page].next[c >> 8]->end[c&0xff] = i | 0x80;
}
return &reverseTables[page];
}
void U32String::decodeOneByte(const char *src, uint32 len, CodePage page) {
const uint16 *conversionTable = getConversionTable(page);
if (conversionTable == nullptr) {
conversionTable = kASCIIConversionTable;
}
ensureCapacity(len, false);
for (uint i = 0; i < len; ++i) {
if ((src[i] & 0x80) == 0) {
operator+=(src[i]);
continue;
}
uint16 val = conversionTable[src[i] & 0x7f];
operator+=(val ? val : invalidCode);
}
}
void String::encodeOneByte(const U32String &src, CodePage page, bool transliterate) {
const ReverseTablePrefixTreeLevel1 *conversionTable =
getReverseConversionTable(page);
ensureCapacity(src.size(), false);
if (conversionTable == nullptr) {
for (uint i = 0; i < src.size(); ++i) {
uint32 c = src[i];
if (c <= 0x7F) {
operator+=((char)c);
continue;
}
if (transliterate) {
translitChar(c);
} else
operator+=('?');
}
return;
}
for (uint i = 0; i < src.size(); ++i) {
uint32 c = src[i];
if (c <= 0x7F) {
operator+=((char)c);
continue;
}
if (c >= kMaxCharSingleByte)
continue;
ReverseTablePrefixTreeLevel2 *l2 = conversionTable->next[c>>8];
unsigned char uc = l2 ? l2->end[c&0xff] : 0;
if (uc != 0) {
operator+=((char)uc);
continue;
}
if (transliterate) {
translitChar(c);
} else
operator+=('?');
}
}
void String::encodeInternal(const U32String &src, CodePage page) {
switch(page) {
case kUtf8:
encodeUTF8(src);
break;
case kWindows932:
encodeWindows932(src);
break;
case kWindows949:
encodeWindows949(src);
break;
case kWindows950:
encodeWindows950(src);
break;
default:
encodeOneByte(src, page);
break;
}
}
U32String convertToU32String(const char *str, CodePage page) {
return String(str).decode(page);
}
U32String convertUtf8ToUtf32(const String &str) {
return str.decode(kUtf8);
}
String convertFromU32String(const U32String &string, CodePage page) {
return string.encode(page);
}
String convertUtf32ToUtf8(const U32String &u32str) {
return u32str.encode(kUtf8);
}
void U32String::decodeInternal(const char *str, uint32 len, CodePage page) {
assert(str);
_storage[0] = 0;
_size = 0;
switch(page) {
case kUtf8:
decodeUTF8(str, len);
break;
case kWindows932:
decodeWindows932(str, len);
break;
case kWindows949:
decodeWindows949(str, len);
break;
case kWindows950:
decodeWindows950(str, len);
break;
default:
decodeOneByte(str, len, page);
break;
}
}
U32String String::decode(CodePage page) const {
if (page == kCodePageInvalid ||
page > kLastEncoding) {
error("Invalid codepage");
}
U32String unicodeString;
unicodeString.decodeInternal(_str, _size, page);
return unicodeString;
}
String U32String::encode(CodePage page) const {
if (page == kCodePageInvalid ||
page > kLastEncoding) {
error("Invalid codepage");
}
String string;
string.encodeInternal(*this, page);
return string;
}
} // End of namespace Common