123 124 static final short ShiftJISEncoding = 2; 125 static final short GBKEncoding = 3; 126 static final short Big5Encoding = 4; 127 static final short WansungEncoding = 5; 128 static final short JohabEncoding = 6; 129 static final short MSUnicodeSurrogateEncoding = 10; 130 131 static final char noSuchChar = (char)0xfffd; 132 static final int SHORTMASK = 0x0000ffff; 133 static final int INTMASK = 0xffffffff; 134 135 static final char[][] converterMaps = new char[7][]; 136 137 /* 138 * Unicode->other encoding translation array. A pre-computed look up 139 * which can be shared across all fonts using that encoding. 140 * Using this saves running character coverters repeatedly. 141 */ 142 char[] xlat; 143 144 static CMap initialize(TrueTypeFont font) { 145 146 CMap cmap = null; 147 148 int offset, platformID, encodingID=-1; 149 150 int three0=0, three1=0, three2=0, three3=0, three4=0, three5=0, 151 three6=0, three10=0; 152 boolean threeStar = false; 153 154 ByteBuffer cmapBuffer = font.getTableBuffer(TrueTypeFont.cmapTag); 155 int cmapTableOffset = font.getTableSize(TrueTypeFont.cmapTag); 156 short numberSubTables = cmapBuffer.getShort(2); 157 158 /* locate the offsets of all 3,* (ie Microsoft platform) encodings */ 159 for (int i=0; i<numberSubTables; i++) { 160 cmapBuffer.position(i * 8 + 4); 161 platformID = cmapBuffer.getShort(); 162 if (platformID == 3) { 163 threeStar = true; 164 encodingID = cmapBuffer.getShort(); 165 offset = cmapBuffer.getInt(); 166 switch (encodingID) { 167 case 0: three0 = offset; break; // MS Symbol encoding 168 case 1: three1 = offset; break; // MS Unicode cmap 169 case 2: three2 = offset; break; // ShiftJIS cmap. 170 case 3: three3 = offset; break; // GBK cmap 171 case 4: three4 = offset; break; // Big 5 cmap 172 case 5: three5 = offset; break; // Wansung 173 case 6: three6 = offset; break; // Johab 174 case 10: three10 = offset; break; // MS Unicode surrogates 175 } 176 } 177 } 178 179 /* This defines the preference order for cmap subtables */ 180 if (threeStar) { 181 if (three10 != 0) { 182 cmap = createCMap(cmapBuffer, three10, null); 183 } 184 else if (three0 != 0) { 185 /* The special case treatment of these fonts leads to 186 * anomalies where a user can view "wingdings" and "wingdings2" 187 * and the latter shows all its code points in the unicode 188 * private use area at 0xF000->0XF0FF and the former shows 189 * a scattered subset of its glyphs that are known mappings to 190 * unicode code points. 191 * The primary purpose of these mappings was to facilitate 192 * display of symbol chars etc in composite fonts, however 193 * this is not needed as all these code points are covered 194 * by Lucida Sans Regular. 195 * Commenting this out reduces the role of these two files 245 else { 246 cmap = createCMap(cmapBuffer, three4, 247 getConverterMap(Big5Encoding)); 248 } 249 } 250 else if (three5 != 0) { 251 cmap = createCMap(cmapBuffer, three5, 252 getConverterMap(WansungEncoding)); 253 } 254 else if (three6 != 0) { 255 cmap = createCMap(cmapBuffer, three6, 256 getConverterMap(JohabEncoding)); 257 } 258 } else { 259 /* No 3,* subtable was found. Just use whatever is the first 260 * table listed. Not very useful but maybe better than 261 * rejecting the font entirely? 262 */ 263 cmap = createCMap(cmapBuffer, cmapBuffer.getInt(8), null); 264 } 265 return cmap; 266 } 267 268 /* speed up the converting by setting the range for double 269 * byte characters; 270 */ 271 static char[] getConverter(short encodingID) { 272 int dBegin = 0x8000; 273 int dEnd = 0xffff; 274 String encoding; 275 276 switch (encodingID) { 277 case ShiftJISEncoding: 278 dBegin = 0x8140; 279 dEnd = 0xfcfc; 280 encoding = "SJIS"; 281 break; 282 case GBKEncoding: 283 dBegin = 0x8140; 284 dEnd = 0xfea0; 407 subtableLength = buffer.getInt(offset+4) & INTMASK; 408 } 409 if (offset+subtableLength > buffer.capacity()) { 410 if (FontUtilities.isLogging()) { 411 FontUtilities.getLogger().warning("Cmap subtable overflows buffer."); 412 } 413 } 414 switch (subtableFormat) { 415 case 0: return new CMapFormat0(buffer, offset); 416 case 2: return new CMapFormat2(buffer, offset, xlat); 417 case 4: return new CMapFormat4(buffer, offset, xlat); 418 case 6: return new CMapFormat6(buffer, offset, xlat); 419 case 8: return new CMapFormat8(buffer, offset, xlat); 420 case 10: return new CMapFormat10(buffer, offset, xlat); 421 case 12: return new CMapFormat12(buffer, offset, xlat); 422 default: throw new RuntimeException("Cmap format unimplemented: " + 423 (int)buffer.getChar(offset)); 424 } 425 } 426 427 /* 428 final char charVal(byte[] cmap, int index) { 429 return (char)(((0xff & cmap[index]) << 8)+(0xff & cmap[index+1])); 430 } 431 432 final short shortVal(byte[] cmap, int index) { 433 return (short)(((0xff & cmap[index]) << 8)+(0xff & cmap[index+1])); 434 } 435 */ 436 abstract char getGlyph(int charCode); 437 438 /* Format 4 Header is 439 * ushort format (off=0) 440 * ushort length (off=2) 441 * ushort language (off=4) 442 * ushort segCountX2 (off=6) 443 * ushort searchRange (off=8) 444 * ushort entrySelector (off=10) 445 * ushort rangeShift (off=12) 446 * ushort endCount[segCount] (off=14) 1042 public static final NullCMapClass theNullCmap = new NullCMapClass(); 1043 1044 final int getControlCodeGlyph(int charCode, boolean noSurrogates) { 1045 if (charCode < 0x0010) { 1046 switch (charCode) { 1047 case 0x0009: 1048 case 0x000a: 1049 case 0x000d: return CharToGlyphMapper.INVISIBLE_GLYPH_ID; 1050 } 1051 } else if (charCode >= 0x200c) { 1052 if ((charCode <= 0x200f) || 1053 (charCode >= 0x2028 && charCode <= 0x202e) || 1054 (charCode >= 0x206a && charCode <= 0x206f)) { 1055 return CharToGlyphMapper.INVISIBLE_GLYPH_ID; 1056 } else if (noSurrogates && charCode >= 0xFFFF) { 1057 return 0; 1058 } 1059 } 1060 return -1; 1061 } 1062 } | 123 124 static final short ShiftJISEncoding = 2; 125 static final short GBKEncoding = 3; 126 static final short Big5Encoding = 4; 127 static final short WansungEncoding = 5; 128 static final short JohabEncoding = 6; 129 static final short MSUnicodeSurrogateEncoding = 10; 130 131 static final char noSuchChar = (char)0xfffd; 132 static final int SHORTMASK = 0x0000ffff; 133 static final int INTMASK = 0xffffffff; 134 135 static final char[][] converterMaps = new char[7][]; 136 137 /* 138 * Unicode->other encoding translation array. A pre-computed look up 139 * which can be shared across all fonts using that encoding. 140 * Using this saves running character coverters repeatedly. 141 */ 142 char[] xlat; 143 UVS uvs = null; 144 145 static CMap initialize(TrueTypeFont font) { 146 147 CMap cmap = null; 148 149 int offset, platformID, encodingID=-1; 150 151 int three0=0, three1=0, three2=0, three3=0, three4=0, three5=0, 152 three6=0, three10=0; 153 int zero5 = 0; // for Unicode Variation Sequences 154 boolean threeStar = false; 155 156 ByteBuffer cmapBuffer = font.getTableBuffer(TrueTypeFont.cmapTag); 157 int cmapTableOffset = font.getTableSize(TrueTypeFont.cmapTag); 158 short numberSubTables = cmapBuffer.getShort(2); 159 160 /* locate the offsets of all 3,* (ie Microsoft platform) encodings */ 161 for (int i=0; i<numberSubTables; i++) { 162 cmapBuffer.position(i * 8 + 4); 163 platformID = cmapBuffer.getShort(); 164 if (platformID == 3) { 165 threeStar = true; 166 encodingID = cmapBuffer.getShort(); 167 offset = cmapBuffer.getInt(); 168 switch (encodingID) { 169 case 0: three0 = offset; break; // MS Symbol encoding 170 case 1: three1 = offset; break; // MS Unicode cmap 171 case 2: three2 = offset; break; // ShiftJIS cmap. 172 case 3: three3 = offset; break; // GBK cmap 173 case 4: three4 = offset; break; // Big 5 cmap 174 case 5: three5 = offset; break; // Wansung 175 case 6: three6 = offset; break; // Johab 176 case 10: three10 = offset; break; // MS Unicode surrogates 177 } 178 } else if (platformID == 0) { 179 encodingID = cmapBuffer.getShort(); 180 offset = cmapBuffer.getInt(); 181 if (encodingID == 5) { 182 zero5 = offset; 183 } 184 } 185 } 186 187 /* This defines the preference order for cmap subtables */ 188 if (threeStar) { 189 if (three10 != 0) { 190 cmap = createCMap(cmapBuffer, three10, null); 191 } 192 else if (three0 != 0) { 193 /* The special case treatment of these fonts leads to 194 * anomalies where a user can view "wingdings" and "wingdings2" 195 * and the latter shows all its code points in the unicode 196 * private use area at 0xF000->0XF0FF and the former shows 197 * a scattered subset of its glyphs that are known mappings to 198 * unicode code points. 199 * The primary purpose of these mappings was to facilitate 200 * display of symbol chars etc in composite fonts, however 201 * this is not needed as all these code points are covered 202 * by Lucida Sans Regular. 203 * Commenting this out reduces the role of these two files 253 else { 254 cmap = createCMap(cmapBuffer, three4, 255 getConverterMap(Big5Encoding)); 256 } 257 } 258 else if (three5 != 0) { 259 cmap = createCMap(cmapBuffer, three5, 260 getConverterMap(WansungEncoding)); 261 } 262 else if (three6 != 0) { 263 cmap = createCMap(cmapBuffer, three6, 264 getConverterMap(JohabEncoding)); 265 } 266 } else { 267 /* No 3,* subtable was found. Just use whatever is the first 268 * table listed. Not very useful but maybe better than 269 * rejecting the font entirely? 270 */ 271 cmap = createCMap(cmapBuffer, cmapBuffer.getInt(8), null); 272 } 273 // For Unicode Variation Sequences 274 if (cmap != null && zero5 != 0) { 275 cmap.createUVS(cmapBuffer, zero5); 276 } 277 return cmap; 278 } 279 280 /* speed up the converting by setting the range for double 281 * byte characters; 282 */ 283 static char[] getConverter(short encodingID) { 284 int dBegin = 0x8000; 285 int dEnd = 0xffff; 286 String encoding; 287 288 switch (encodingID) { 289 case ShiftJISEncoding: 290 dBegin = 0x8140; 291 dEnd = 0xfcfc; 292 encoding = "SJIS"; 293 break; 294 case GBKEncoding: 295 dBegin = 0x8140; 296 dEnd = 0xfea0; 419 subtableLength = buffer.getInt(offset+4) & INTMASK; 420 } 421 if (offset+subtableLength > buffer.capacity()) { 422 if (FontUtilities.isLogging()) { 423 FontUtilities.getLogger().warning("Cmap subtable overflows buffer."); 424 } 425 } 426 switch (subtableFormat) { 427 case 0: return new CMapFormat0(buffer, offset); 428 case 2: return new CMapFormat2(buffer, offset, xlat); 429 case 4: return new CMapFormat4(buffer, offset, xlat); 430 case 6: return new CMapFormat6(buffer, offset, xlat); 431 case 8: return new CMapFormat8(buffer, offset, xlat); 432 case 10: return new CMapFormat10(buffer, offset, xlat); 433 case 12: return new CMapFormat12(buffer, offset, xlat); 434 default: throw new RuntimeException("Cmap format unimplemented: " + 435 (int)buffer.getChar(offset)); 436 } 437 } 438 439 private void createUVS(ByteBuffer buffer, int offset) { 440 int subtableFormat = buffer.getChar(offset); 441 if (subtableFormat == 14) { 442 long subtableLength = buffer.getInt(offset + 2) & INTMASK; 443 if (offset + subtableLength > buffer.capacity()) { 444 if (FontUtilities.isLogging()) { 445 FontUtilities.getLogger() 446 .warning("Cmap UVS subtable overflows buffer."); 447 } 448 } 449 try { 450 this.uvs = new UVS(buffer, offset); 451 } catch (Throwable t) { 452 t.printStackTrace(); 453 } 454 } 455 return; 456 } 457 458 /* 459 final char charVal(byte[] cmap, int index) { 460 return (char)(((0xff & cmap[index]) << 8)+(0xff & cmap[index+1])); 461 } 462 463 final short shortVal(byte[] cmap, int index) { 464 return (short)(((0xff & cmap[index]) << 8)+(0xff & cmap[index+1])); 465 } 466 */ 467 abstract char getGlyph(int charCode); 468 469 /* Format 4 Header is 470 * ushort format (off=0) 471 * ushort length (off=2) 472 * ushort language (off=4) 473 * ushort segCountX2 (off=6) 474 * ushort searchRange (off=8) 475 * ushort entrySelector (off=10) 476 * ushort rangeShift (off=12) 477 * ushort endCount[segCount] (off=14) 1073 public static final NullCMapClass theNullCmap = new NullCMapClass(); 1074 1075 final int getControlCodeGlyph(int charCode, boolean noSurrogates) { 1076 if (charCode < 0x0010) { 1077 switch (charCode) { 1078 case 0x0009: 1079 case 0x000a: 1080 case 0x000d: return CharToGlyphMapper.INVISIBLE_GLYPH_ID; 1081 } 1082 } else if (charCode >= 0x200c) { 1083 if ((charCode <= 0x200f) || 1084 (charCode >= 0x2028 && charCode <= 0x202e) || 1085 (charCode >= 0x206a && charCode <= 0x206f)) { 1086 return CharToGlyphMapper.INVISIBLE_GLYPH_ID; 1087 } else if (noSurrogates && charCode >= 0xFFFF) { 1088 return 0; 1089 } 1090 } 1091 return -1; 1092 } 1093 1094 static class UVS { 1095 int numSelectors; 1096 int[] selector; 1097 1098 //for Non-Default UVS Table 1099 int[] numUVSMapping; 1100 int[][] unicodeValue; 1101 char[][] glyphID; 1102 1103 UVS(ByteBuffer buffer, int offset) { 1104 numSelectors = buffer.getInt(offset+6); 1105 selector = new int[numSelectors]; 1106 numUVSMapping = new int[numSelectors]; 1107 unicodeValue = new int[numSelectors][]; 1108 glyphID = new char[numSelectors][]; 1109 1110 for (int i = 0; i < numSelectors; i++) { 1111 buffer.position(offset + 10 + i * 11); 1112 selector[i] = (buffer.get() & 0xff) << 16; //UINT24 1113 selector[i] += (buffer.get() & 0xff) << 8; 1114 selector[i] += buffer.get() & 0xff; 1115 1116 //skip Default UVS Table 1117 1118 //for Non-Default UVS Table 1119 int tableOffset = buffer.getInt(offset + 10 + i * 11 + 7); 1120 if (tableOffset == 0) { 1121 numUVSMapping[i] = 0; 1122 } else if (tableOffset > 0) { 1123 buffer.position(offset+tableOffset); 1124 numUVSMapping[i] = buffer.getInt() & INTMASK; 1125 unicodeValue[i] = new int[numUVSMapping[i]]; 1126 glyphID[i] = new char[numUVSMapping[i]]; 1127 1128 for (int j = 0; j < numUVSMapping[i]; j++) { 1129 int temp = (buffer.get() & 0xff) << 16; //UINT24 1130 temp += (buffer.get() & 0xff) << 8; 1131 temp += buffer.get() & 0xff; 1132 unicodeValue[i][j] = temp; 1133 glyphID[i][j] = buffer.getChar(); 1134 } 1135 } 1136 } 1137 } 1138 1139 static final int VS_NOGLYPH = 0; 1140 private int getGlyph(int charCode, int variationSelector) { 1141 int targetSelector = -1; 1142 for (int i = 0; i < numSelectors; i++) { 1143 if (selector[i] == variationSelector) { 1144 targetSelector = i; 1145 break; 1146 } 1147 } 1148 if (targetSelector == -1) { 1149 return VS_NOGLYPH; 1150 } 1151 if (numUVSMapping[targetSelector] > 0) { 1152 int index = java.util.Arrays.binarySearch( 1153 unicodeValue[targetSelector], charCode); 1154 if (index >= 0) { 1155 return glyphID[targetSelector][index]; 1156 } 1157 } 1158 return VS_NOGLYPH; 1159 } 1160 } 1161 1162 char getVariationGlyph(int charCode, int variationSelector) { 1163 char glyph = 0; 1164 if (uvs == null) { 1165 glyph = getGlyph(charCode); 1166 } else { 1167 int result = uvs.getGlyph(charCode, variationSelector); 1168 if (result > 0) { 1169 glyph = (char)(result & 0xFFFF); 1170 } else { 1171 glyph = getGlyph(charCode); 1172 } 1173 } 1174 return glyph; 1175 } 1176 } |