icuSources/layout/ThaiShaping.cpp

   1 /*
   2  * @(#)ThaiShaping.cpp  1.13 00/03/15
   3  *
   4  * (C) Copyright IBM Corp. 1998-2003 - All Rights Reserved
   5  *
   6  */
   7
   8 #include "LETypes.h"
   9 #include "LEGlyphFilter.h"
  10 #include "OpenTypeTables.h"
  11 #include "ThaiShaping.h"
  12
  13 U_NAMESPACE_BEGIN
  14
  15 enum {
  16     CH_SPACE        = 0x0020,
  17     CH_YAMAKKAN     = 0x0E4E,
  18     CH_MAI_HANAKAT  = 0x0E31,
  19     CH_SARA_AA      = 0x0E32,
  20     CH_SARA_AM      = 0x0E33,
  21     CH_SARA_UEE     = 0x0E37,
  22     CH_MAITAIKHU    = 0x0E47,
  23     CH_NIKHAHIT     = 0x0E4D,
  24     CH_SARA_U       = 0x0E38,
  25     CH_PHINTHU      = 0x0E3A,
  26     CH_YO_YING      = 0x0E0D,
  27     CH_THO_THAN     = 0x0E10,
  28     CH_DOTTED_CIRCLE = 0x25CC
  29 };
  30
  31     le_uint8 ThaiShaping::getCharClass(LEUnicode ch)
  32 {
  33     le_uint8 charClass = NON;
  34
  35     if (ch >= 0x0E00 && ch <= 0x0E5B) {
  36         charClass = classTable[ch - 0x0E00];
  37     }
  38
  39     return charClass;
  40 }
  41
  42
  43 LEUnicode ThaiShaping::leftAboveVowel(LEUnicode vowel, le_uint8 glyphSet)
  44 {
  45     static LEUnicode leftAboveVowels[][7] = {
  46         {0x0E61, 0x0E32, 0x0E33, 0x0E64, 0x0E65, 0x0E66, 0x0E67},
  47         {0xF710, 0x0E32, 0x0E33, 0xF701, 0xF702, 0xF703, 0xF704},
  48         {0xF884, 0x0E32, 0x0E33, 0xF885, 0xF886, 0xF887, 0xF788},
  49         {0x0E31, 0x0E32, 0x0E33, 0x0E34, 0x0E35, 0x0E36, 0x0E37}
  50     };
  51
  52     if (vowel >= CH_MAI_HANAKAT && vowel <= CH_SARA_UEE) {
  53         return leftAboveVowels[glyphSet][vowel - CH_MAI_HANAKAT];
  54     }
  55
  56     if (vowel == CH_YAMAKKAN && glyphSet == 0) {
  57         return 0x0E7E;
  58     }
  59
  60     return vowel;
  61 }
  62
  63 LEUnicode ThaiShaping::lowerRightTone(LEUnicode tone, le_uint8 glyphSet)
  64 {
  65     static LEUnicode lowerRightTones[][7] = {
  66         {0x0E68, 0x0E69, 0x0E6A, 0x0E6B, 0x0E6C, 0x0E6D, 0x0E6E},
  67         {0x0E47, 0xF70A, 0xF70B, 0xF70C, 0xF70D, 0xF70E, 0x0E4D},
  68         {0x0E47, 0xF88B, 0xF88E, 0xF891, 0xF894, 0xF897, 0x0E4D},
  69         {0x0E47, 0x0E48, 0x0E49, 0x0E4A, 0x0E4B, 0x0E4C, 0x0E4D}
  70     };
  71
  72     if (tone >= CH_MAITAIKHU && tone <= CH_NIKHAHIT) {
  73         return lowerRightTones[glyphSet][tone - CH_MAITAIKHU];
  74     }
  75
  76     return tone;
  77 }
  78
  79 LEUnicode ThaiShaping::lowerLeftTone(LEUnicode tone, le_uint8 glyphSet)
  80 {
  81     static LEUnicode lowerLeftTones[][7] = {
  82         {0x0E76, 0x0E77, 0x0E78, 0x0E79, 0x0E7A, 0x0E7B, 0x0E7C},
  83         {0xF712, 0xF705, 0xF706, 0xF707, 0xF708, 0xF709, 0xF711},
  84         {0xF889, 0xF88C, 0xF88F, 0xF892, 0xF895, 0xF898, 0xF899},
  85         {0x0E47, 0x0E48, 0x0E49, 0x0E4A, 0x0E4B, 0x0E4C, 0x0E4D}
  86     };
  87
  88     if (tone >= CH_MAITAIKHU && tone <= CH_NIKHAHIT) {
  89         return lowerLeftTones[glyphSet][tone - CH_MAITAIKHU];
  90     }
  91
  92     return tone;
  93 }
  94
  95 LEUnicode ThaiShaping::upperLeftTone(LEUnicode tone, le_uint8 glyphSet)
  96 {
  97     static LEUnicode upperLeftTones[][7] = {
  98         {0x0E6F, 0x0E70, 0x0E71, 0x0E72, 0x0E73, 0x0E74, 0x0E75},
  99         {0xF712, 0xF713, 0xF714, 0xF715, 0xF716, 0xF717, 0xF711},
 100         {0xF889, 0xF88A, 0xF88D, 0xF890, 0xF893, 0xF896, 0xF899},
 101         {0x0E47, 0x0E48, 0x0E49, 0x0E4A, 0x0E4B, 0x0E4C, 0x0E4D}
 102     };
 103
 104     if (tone >= CH_MAITAIKHU && tone <= CH_NIKHAHIT) {
 105         return upperLeftTones[glyphSet][tone - CH_MAITAIKHU];
 106     }
 107
 108     return tone;
 109 }
 110
 111 LEUnicode ThaiShaping::lowerBelowVowel(LEUnicode vowel, le_uint8 glyphSet)
 112 {
 113     static LEUnicode lowerBelowVowels[][3] = {
 114         {0x0E3C, 0x0E3D, 0x0E3E},
 115         {0xF718, 0xF719, 0xF71A},
 116         {0x0E38, 0x0E39, 0x0E3A},
 117         {0x0E38, 0x0E39, 0x0E3A}
 118
 119     };
 120
 121     if (vowel >= CH_SARA_U && vowel <= CH_PHINTHU) {
 122         return lowerBelowVowels[glyphSet][vowel - CH_SARA_U];
 123     }
 124
 125     return vowel;
 126 }
 127
 128 LEUnicode ThaiShaping::noDescenderCOD(LEUnicode cod, le_uint8 glyphSet)
 129 {
 130     static LEUnicode noDescenderCODs[][4] = {
 131         {0x0E60, 0x0E0E, 0x0E0F, 0x0E63},
 132         {0xF70F, 0x0E0E, 0x0E0F, 0xF700},
 133         {0x0E0D, 0x0E0E, 0x0E0F, 0x0E10},
 134         {0x0E0D, 0x0E0E, 0x0E0F, 0x0E10}
 135
 136     };
 137
 138     if (cod >= CH_YO_YING && cod <= CH_THO_THAN) {
 139         return noDescenderCODs[glyphSet][cod - CH_YO_YING];
 140     }
 141
 142     return cod;
 143 }
 144
 145 le_uint8 ThaiShaping::doTransition (StateTransition transition, LEUnicode currChar, le_int32 inputIndex, le_uint8 glyphSet,
 146         LEUnicode errorChar, LEUnicode *outputBuffer, le_int32 *charIndicies, le_int32 &outputIndex)
 147 {
 148     switch (transition.action) {
 149     case _A:
 150         charIndicies[outputIndex] = inputIndex;
 151         outputBuffer[outputIndex++] = currChar;
 152         break;
 153
 154     case _C:
 155         charIndicies[outputIndex] = inputIndex;
 156         outputBuffer[outputIndex++] = currChar;
 157         break;
 158
 159     case _D:
 160         charIndicies[outputIndex] = inputIndex;
 161         outputBuffer[outputIndex++] = leftAboveVowel(currChar, glyphSet);
 162         break;
 163
 164     case _E:
 165         charIndicies[outputIndex] = inputIndex;
 166         outputBuffer[outputIndex++] = lowerRightTone(currChar, glyphSet);
 167         break;
 168
 169     case _F:
 170         charIndicies[outputIndex] = inputIndex;
 171         outputBuffer[outputIndex++] = lowerLeftTone(currChar, glyphSet);
 172         break;
 173
 174     case _G:
 175         charIndicies[outputIndex] = inputIndex;
 176         outputBuffer[outputIndex++] = upperLeftTone(currChar, glyphSet);
 177         break;
 178
 179     case _H:
 180     {
 181         LEUnicode cod = outputBuffer[outputIndex - 1];
 182         LEUnicode coa = noDescenderCOD(cod, glyphSet);
 183
 184         if (cod != coa) {
 185             outputBuffer[outputIndex - 1] = coa;
 186
 187             charIndicies[outputIndex] = inputIndex;
 188             outputBuffer[outputIndex++] = currChar;
 189             break;
 190         }
 191
 192         charIndicies[outputIndex] = inputIndex;
 193         outputBuffer[outputIndex++] = lowerBelowVowel(currChar, glyphSet);
 194         break;
 195     }
 196
 197     case _R:
 198         charIndicies[outputIndex] = inputIndex;
 199         outputBuffer[outputIndex++] = errorChar;
 200
 201         charIndicies[outputIndex] = inputIndex;
 202         outputBuffer[outputIndex++] = currChar;
 203         break;
 204
 205     case _S:
 206         if (currChar == CH_SARA_AM) {
 207             charIndicies[outputIndex] = inputIndex;
 208             outputBuffer[outputIndex++] = errorChar;
 209         }
 210
 211         charIndicies[outputIndex] = inputIndex;
 212         outputBuffer[outputIndex++] = currChar;
 213         break;
 214
 215     default:
 216         // FIXME: if we get here, there's an error
 217         // in the state table!
 218         charIndicies[outputIndex] = inputIndex;
 219         outputBuffer[outputIndex++] = currChar;
 220         break;
 221      }
 222
 223      return transition.nextState;
 224 }
 225
 226 le_uint8 ThaiShaping::getNextState(LEUnicode ch, le_uint8 prevState, le_int32 inputIndex, le_uint8 glyphSet, LEUnicode errorChar,
 227                               le_uint8 &charClass, LEUnicode *output, le_int32 *charIndicies, le_int32 &outputIndex)
 228 {
 229     StateTransition transition;
 230
 231     charClass = getCharClass(ch);
 232     transition = getTransition(prevState, charClass);
 233
 234     return doTransition(transition, ch, inputIndex, glyphSet, errorChar, output, charIndicies, outputIndex);
 235 }
 236
 237 le_bool ThaiShaping::isLegalHere(LEUnicode ch, le_uint8 prevState)
 238 {
 239     le_uint8 charClass = getCharClass(ch);
 240     StateTransition transition = getTransition(prevState, charClass);
 241
 242     switch (transition.action) {
 243     case _A:
 244     case _C:
 245     case _D:
 246     case _E:
 247     case _F:
 248     case _G:
 249     case _H:
 250         return true;
 251
 252     case _R:
 253     case _S:
 254         return false;
 255
 256     default:
 257         // FIXME: if we get here, there's an error
 258         // in the state table!
 259         return false;
 260     }
 261 }
 262
 263 le_int32 ThaiShaping::compose(const LEUnicode *input, le_int32 offset, le_int32 charCount, le_uint8 glyphSet,
 264                           LEUnicode errorChar, LEUnicode *output, le_int32 *charIndicies)
 265 {
 266     le_uint8 state = 0;
 267     le_int32 inputIndex;
 268     le_int32 outputIndex = 0;
 269     le_uint8 conState = 0xFF;
 270     le_int32 conInput = -1;
 271     le_int32 conOutput = -1;
 272
 273     for (inputIndex = 0; inputIndex < charCount; inputIndex += 1) {
 274         LEUnicode ch = input[inputIndex + offset];
 275         le_uint8 charClass;
 276
 277         // Decompose SARA AM into NIKHAHIT + SARA AA
 278         if (ch == CH_SARA_AM && isLegalHere(ch, state)) {
 279             outputIndex = conOutput;
 280             state = getNextState(CH_NIKHAHIT, conState, inputIndex, glyphSet, errorChar, charClass,
 281                 output, charIndicies, outputIndex);
 282
 283             for (int j = conInput + 1; j < inputIndex; j += 1) {
 284                 ch = input[j + offset];
 285                 state = getNextState(ch, state, j, glyphSet, errorChar, charClass,
 286                     output, charIndicies, outputIndex);
 287             }
 288
 289             ch = CH_SARA_AA;
 290         }
 291
 292         state = getNextState(ch, state, inputIndex, glyphSet, errorChar, charClass,
 293             output, charIndicies, outputIndex);
 294
 295         if (charClass >= CON && charClass <= COD) {
 296             conState = state;
 297             conInput = inputIndex;
 298             conOutput = outputIndex;
 299         }
 300     }
 301
 302     return outputIndex;
 303 }
 304
 305 U_NAMESPACE_END