wtf/unicode/UTF8.cpp

   1 /*
   2  * Copyright (C) 2007 Apple Inc.  All rights reserved.
   3  *
   4  * Redistribution and use in source and binary forms, with or without
   5  * modification, are permitted provided that the following conditions
   6  * are met:
   7  * 1. Redistributions of source code must retain the above copyright
   8  *    notice, this list of conditions and the following disclaimer.
   9  * 2. Redistributions in binary form must reproduce the above copyright
  10  *    notice, this list of conditions and the following disclaimer in the
  11  *    documentation and/or other materials provided with the distribution.
  12  *
  13  * THIS SOFTWARE IS PROVIDED BY APPLE COMPUTER, INC. ``AS IS'' AND ANY
  14  * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  15  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
  16  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL APPLE COMPUTER, INC. OR
  17  * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
  18  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
  19  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
  20  * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
  21  * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  22  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
  23  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  24  */
  25
  26 #include "config.h"
  27 #include "UTF8.h"
  28
  29 namespace WTF {
  30 namespace Unicode {
  31
  32 inline int inlineUTF8SequenceLengthNonASCII(char b0)
  33 {
  34     if ((b0 & 0xC0) != 0xC0)
  35         return 0;
  36     if ((b0 & 0xE0) == 0xC0)
  37         return 2;
  38     if ((b0 & 0xF0) == 0xE0)
  39         return 3;
  40     if ((b0 & 0xF8) == 0xF0)
  41         return 4;
  42     return 0;
  43 }
  44
  45 inline int inlineUTF8SequenceLength(char b0)
  46 {
  47     return (b0 & 0x80) == 0 ? 1 : inlineUTF8SequenceLengthNonASCII(b0);
  48 }
  49
  50 int UTF8SequenceLength(char b0)
  51 {
  52     return (b0 & 0x80) == 0 ? 1 : inlineUTF8SequenceLengthNonASCII(b0);
  53 }
  54
  55 int decodeUTF8Sequence(const char* sequence)
  56 {
  57     // Handle 0-byte sequences (never valid).
  58     const unsigned char b0 = sequence[0];
  59     const int length = inlineUTF8SequenceLength(b0);
  60     if (length == 0)
  61         return -1;
  62
  63     // Handle 1-byte sequences (plain ASCII).
  64     const unsigned char b1 = sequence[1];
  65     if (length == 1) {
  66         if (b1)
  67             return -1;
  68         return b0;
  69     }
  70
  71     // Handle 2-byte sequences.
  72     if ((b1 & 0xC0) != 0x80)
  73         return -1;
  74     const unsigned char b2 = sequence[2];
  75     if (length == 2) {
  76         if (b2)
  77             return -1;
  78         const int c = ((b0 & 0x1F) << 6) | (b1 & 0x3F);
  79         if (c < 0x80)
  80             return -1;
  81         return c;
  82     }
  83
  84     // Handle 3-byte sequences.
  85     if ((b2 & 0xC0) != 0x80)
  86         return -1;
  87     const unsigned char b3 = sequence[3];
  88     if (length == 3) {
  89         if (b3)
  90             return -1;
  91         const int c = ((b0 & 0xF) << 12) | ((b1 & 0x3F) << 6) | (b2 & 0x3F);
  92         if (c < 0x800)
  93             return -1;
  94         // UTF-16 surrogates should never appear in UTF-8 data.
  95         if (c >= 0xD800 && c <= 0xDFFF)
  96             return -1;
  97         return c;
  98     }
  99
 100     // Handle 4-byte sequences.
 101     if ((b3 & 0xC0) != 0x80)
 102         return -1;
 103     const unsigned char b4 = sequence[4];
 104     if (length == 4) {
 105         if (b4)
 106             return -1;
 107         const int c = ((b0 & 0x7) << 18) | ((b1 & 0x3F) << 12) | ((b2 & 0x3F) << 6) | (b3 & 0x3F);
 108         if (c < 0x10000 || c > 0x10FFFF)
 109             return -1;
 110         return c;
 111     }
 112
 113     return -1;
 114 }
 115
 116 // Once the bits are split out into bytes of UTF-8, this is a mask OR-ed
 117 // into the first byte, depending on how many bytes follow.  There are
 118 // as many entries in this table as there are UTF-8 sequence types.
 119 // (I.e., one byte sequence, two byte... etc.). Remember that sequencs
 120 // for *legal* UTF-8 will be 4 or fewer bytes total.
 121 static const unsigned char firstByteMark[7] = { 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
 122
 123 ConversionResult convertUTF16ToUTF8(
 124     const UChar** sourceStart, const UChar* sourceEnd,
 125     char** targetStart, char* targetEnd, bool strict)
 126 {
 127     ConversionResult result = conversionOK;
 128     const UChar* source = *sourceStart;
 129     char* target = *targetStart;
 130     while (source < sourceEnd) {
 131         UChar32 ch;
 132         unsigned short bytesToWrite = 0;
 133         const UChar32 byteMask = 0xBF;
 134         const UChar32 byteMark = 0x80;
 135         const UChar* oldSource = source; // In case we have to back up because of target overflow.
 136         ch = static_cast<unsigned short>(*source++);
 137         // If we have a surrogate pair, convert to UChar32 first.
 138         if (ch >= 0xD800 && ch <= 0xDBFF) {
 139             // If the 16 bits following the high surrogate are in the source buffer...
 140             if (source < sourceEnd) {
 141                 UChar32 ch2 = static_cast<unsigned short>(*source);
 142                 // If it's a low surrogate, convert to UChar32.
 143                 if (ch2 >= 0xDC00 && ch2 <= 0xDFFF) {
 144                     ch = ((ch - 0xD800) << 10) + (ch2 - 0xDC00) + 0x0010000;
 145                     ++source;
 146                 } else if (strict) { // it's an unpaired high surrogate
 147                     --source; // return to the illegal value itself
 148                     result = sourceIllegal;
 149                     break;
 150                 }
 151             } else { // We don't have the 16 bits following the high surrogate.
 152                 --source; // return to the high surrogate
 153                 result = sourceExhausted;
 154                 break;
 155             }
 156         } else if (strict) {
 157             // UTF-16 surrogate values are illegal in UTF-32
 158             if (ch >= 0xDC00 && ch <= 0xDFFF) {
 159                 --source; // return to the illegal value itself
 160                 result = sourceIllegal;
 161                 break;
 162             }
 163         }
 164         // Figure out how many bytes the result will require
 165         if (ch < (UChar32)0x80) {
 166             bytesToWrite = 1;
 167         } else if (ch < (UChar32)0x800) {
 168             bytesToWrite = 2;
 169         } else if (ch < (UChar32)0x10000) {
 170             bytesToWrite = 3;
 171         } else if (ch < (UChar32)0x110000) {
 172             bytesToWrite = 4;
 173         } else {
 174             bytesToWrite = 3;
 175             ch = 0xFFFD;
 176         }
 177
 178         target += bytesToWrite;
 179         if (target > targetEnd) {
 180             source = oldSource; // Back up source pointer!
 181             target -= bytesToWrite;
 182             result = targetExhausted;
 183             break;
 184         }
 185         switch (bytesToWrite) { // note: everything falls through.
 186             case 4: *--target = (char)((ch | byteMark) & byteMask); ch >>= 6;
 187             case 3: *--target = (char)((ch | byteMark) & byteMask); ch >>= 6;
 188             case 2: *--target = (char)((ch | byteMark) & byteMask); ch >>= 6;
 189             case 1: *--target =  (char)(ch | firstByteMark[bytesToWrite]);
 190         }
 191         target += bytesToWrite;
 192     }
 193     *sourceStart = source;
 194     *targetStart = target;
 195     return result;
 196 }
 197
 198 // This must be called with the length pre-determined by the first byte.
 199 // If presented with a length > 4, this returns false.  The Unicode
 200 // definition of UTF-8 goes up to 4-byte sequences.
 201 static bool isLegalUTF8(const unsigned char* source, int length)
 202 {
 203     unsigned char a;
 204     const unsigned char* srcptr = source + length;
 205     switch (length) {
 206         default: return false;
 207         // Everything else falls through when "true"...
 208         case 4: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
 209         case 3: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
 210         case 2: if ((a = (*--srcptr)) > 0xBF) return false;
 211
 212         switch (*source) {
 213             // no fall-through in this inner switch
 214             case 0xE0: if (a < 0xA0) return false; break;
 215             case 0xED: if (a > 0x9F) return false; break;
 216             case 0xF0: if (a < 0x90) return false; break;
 217             case 0xF4: if (a > 0x8F) return false; break;
 218             default:   if (a < 0x80) return false;
 219         }
 220
 221         case 1: if (*source >= 0x80 && *source < 0xC2) return false;
 222     }
 223     if (*source > 0xF4)
 224         return false;
 225     return true;
 226 }
 227
 228 // Magic values subtracted from a buffer value during UTF8 conversion.
 229 // This table contains as many values as there might be trailing bytes
 230 // in a UTF-8 sequence.
 231 static const UChar32 offsetsFromUTF8[6] = { 0x00000000UL, 0x00003080UL, 0x000E2080UL,
 232             0x03C82080UL, 0xFA082080UL, 0x82082080UL };
 233
 234 ConversionResult convertUTF8ToUTF16(
 235     const char** sourceStart, const char* sourceEnd,
 236     UChar** targetStart, UChar* targetEnd, bool strict)
 237 {
 238     ConversionResult result = conversionOK;
 239     const char* source = *sourceStart;
 240     UChar* target = *targetStart;
 241     while (source < sourceEnd) {
 242         UChar32 ch = 0;
 243         int extraBytesToRead = UTF8SequenceLength(*source) - 1;
 244         if (source + extraBytesToRead >= sourceEnd) {
 245             result = sourceExhausted;
 246             break;
 247         }
 248         // Do this check whether lenient or strict
 249         if (!isLegalUTF8(reinterpret_cast<const unsigned char*>(source), extraBytesToRead + 1)) {
 250             result = sourceIllegal;
 251             break;
 252         }
 253         // The cases all fall through.
 254         switch (extraBytesToRead) {
 255             case 5: ch += static_cast<unsigned char>(*source++); ch <<= 6; // remember, illegal UTF-8
 256             case 4: ch += static_cast<unsigned char>(*source++); ch <<= 6; // remember, illegal UTF-8
 257             case 3: ch += static_cast<unsigned char>(*source++); ch <<= 6;
 258             case 2: ch += static_cast<unsigned char>(*source++); ch <<= 6;
 259             case 1: ch += static_cast<unsigned char>(*source++); ch <<= 6;
 260             case 0: ch += static_cast<unsigned char>(*source++);
 261         }
 262         ch -= offsetsFromUTF8[extraBytesToRead];
 263
 264         if (target >= targetEnd) {
 265             source -= (extraBytesToRead + 1); // Back up source pointer!
 266             result = targetExhausted; break;
 267         }
 268         if (ch <= 0xFFFF) {
 269             // UTF-16 surrogate values are illegal in UTF-32
 270             if (ch >= 0xD800 && ch <= 0xDFFF) {
 271                 if (strict) {
 272                     source -= (extraBytesToRead + 1); // return to the illegal value itself
 273                     result = sourceIllegal;
 274                     break;
 275                 } else
 276                     *target++ = 0xFFFD;
 277             } else
 278                 *target++ = (UChar)ch; // normal case
 279         } else if (ch > 0x10FFFF) {
 280             if (strict) {
 281                 result = sourceIllegal;
 282                 source -= (extraBytesToRead + 1); // return to the start
 283                 break; // Bail out; shouldn't continue
 284             } else
 285                 *target++ = 0xFFFD;
 286         } else {
 287             // target is a character in range 0xFFFF - 0x10FFFF
 288             if (target + 1 >= targetEnd) {
 289                 source -= (extraBytesToRead + 1); // Back up source pointer!
 290                 result = targetExhausted;
 291                 break;
 292             }
 293             ch -= 0x0010000UL;
 294             *target++ = (UChar)((ch >> 10) + 0xD800);
 295             *target++ = (UChar)((ch & 0x03FF) + 0xDC00);
 296         }
 297     }
 298     *sourceStart = source;
 299     *targetStart = target;
 300     return result;
 301 }
 302
 303 }
 304 }