| 1 | // © 2016 and later: Unicode, Inc. and others. |
| 2 | // License & terms of use: http://www.unicode.org/copyright.html |
| 3 | /* |
| 4 | ********************************************************************** |
| 5 | * Copyright (C) 1999-2016, International Business Machines |
| 6 | * Corporation and others. All Rights Reserved. |
| 7 | ********************************************************************** |
| 8 | * |
| 9 | * File USC_IMPL.C |
| 10 | * |
| 11 | * Modification History: |
| 12 | * |
| 13 | * Date Name Description |
| 14 | * 07/08/2002 Eric Mader Creation. |
| 15 | ****************************************************************************** |
| 16 | */ |
| 17 | |
| 18 | #include "unicode/uscript.h" |
| 19 | #include "usc_impl.h" |
| 20 | #include "cmemory.h" |
| 21 | |
| 22 | #define PAREN_STACK_DEPTH 32 |
| 23 | |
| 24 | #define MOD(sp) ((sp) % PAREN_STACK_DEPTH) |
| 25 | #define LIMIT_INC(sp) (((sp) < PAREN_STACK_DEPTH)? (sp) + 1 : PAREN_STACK_DEPTH) |
| 26 | #define INC(sp,count) (MOD((sp) + (count))) |
| 27 | #define INC1(sp) (INC(sp, 1)) |
| 28 | #define DEC(sp,count) (MOD((sp) + PAREN_STACK_DEPTH - (count))) |
| 29 | #define DEC1(sp) (DEC(sp, 1)) |
| 30 | #define STACK_IS_EMPTY(scriptRun) ((scriptRun)->pushCount <= 0) |
| 31 | #define STACK_IS_NOT_EMPTY(scriptRun) (! STACK_IS_EMPTY(scriptRun)) |
| 32 | #define TOP(scriptRun) ((scriptRun)->parenStack[(scriptRun)->parenSP]) |
| 33 | #define SYNC_FIXUP(scriptRun) ((scriptRun)->fixupCount = 0) |
| 34 | |
| 35 | struct ParenStackEntry |
| 36 | { |
| 37 | int32_t pairIndex; |
| 38 | UScriptCode scriptCode; |
| 39 | }; |
| 40 | |
| 41 | struct UScriptRun |
| 42 | { |
| 43 | int32_t textLength; |
| 44 | const UChar *textArray; |
| 45 | |
| 46 | int32_t scriptStart; |
| 47 | int32_t scriptLimit; |
| 48 | UScriptCode scriptCode; |
| 49 | |
| 50 | struct ParenStackEntry parenStack[PAREN_STACK_DEPTH]; |
| 51 | int32_t parenSP; |
| 52 | int32_t pushCount; |
| 53 | int32_t fixupCount; |
| 54 | }; |
| 55 | |
| 56 | static int8_t highBit(int32_t value); |
| 57 | |
| 58 | static const UChar32 pairedChars[] = { |
| 59 | 0x0028, 0x0029, /* ascii paired punctuation */ |
| 60 | 0x003c, 0x003e, |
| 61 | 0x005b, 0x005d, |
| 62 | 0x007b, 0x007d, |
| 63 | 0x00ab, 0x00bb, /* guillemets */ |
| 64 | 0x2018, 0x2019, /* general punctuation */ |
| 65 | 0x201c, 0x201d, |
| 66 | 0x2039, 0x203a, |
| 67 | 0x3008, 0x3009, /* chinese paired punctuation */ |
| 68 | 0x300a, 0x300b, |
| 69 | 0x300c, 0x300d, |
| 70 | 0x300e, 0x300f, |
| 71 | 0x3010, 0x3011, |
| 72 | 0x3014, 0x3015, |
| 73 | 0x3016, 0x3017, |
| 74 | 0x3018, 0x3019, |
| 75 | 0x301a, 0x301b |
| 76 | }; |
| 77 | |
| 78 | static void push(UScriptRun *scriptRun, int32_t pairIndex, UScriptCode scriptCode) |
| 79 | { |
| 80 | scriptRun->pushCount = LIMIT_INC(scriptRun->pushCount); |
| 81 | scriptRun->fixupCount = LIMIT_INC(scriptRun->fixupCount); |
| 82 | |
| 83 | scriptRun->parenSP = INC1(scriptRun->parenSP); |
| 84 | scriptRun->parenStack[scriptRun->parenSP].pairIndex = pairIndex; |
| 85 | scriptRun->parenStack[scriptRun->parenSP].scriptCode = scriptCode; |
| 86 | } |
| 87 | |
| 88 | static void pop(UScriptRun *scriptRun) |
| 89 | { |
| 90 | if (STACK_IS_EMPTY(scriptRun)) { |
| 91 | return; |
| 92 | } |
| 93 | |
| 94 | if (scriptRun->fixupCount > 0) { |
| 95 | scriptRun->fixupCount -= 1; |
| 96 | } |
| 97 | |
| 98 | scriptRun->pushCount -= 1; |
| 99 | scriptRun->parenSP = DEC1(scriptRun->parenSP); |
| 100 | |
| 101 | /* If the stack is now empty, reset the stack |
| 102 | pointers to their initial values. |
| 103 | */ |
| 104 | if (STACK_IS_EMPTY(scriptRun)) { |
| 105 | scriptRun->parenSP = -1; |
| 106 | } |
| 107 | } |
| 108 | |
| 109 | static void fixup(UScriptRun *scriptRun, UScriptCode scriptCode) |
| 110 | { |
| 111 | int32_t fixupSP = DEC(scriptRun->parenSP, scriptRun->fixupCount); |
| 112 | |
| 113 | while (scriptRun->fixupCount-- > 0) { |
| 114 | fixupSP = INC1(fixupSP); |
| 115 | scriptRun->parenStack[fixupSP].scriptCode = scriptCode; |
| 116 | } |
| 117 | } |
| 118 | |
| 119 | static int8_t |
| 120 | highBit(int32_t value) |
| 121 | { |
| 122 | int8_t bit = 0; |
| 123 | |
| 124 | if (value <= 0) { |
| 125 | return -32; |
| 126 | } |
| 127 | |
| 128 | if (value >= 1 << 16) { |
| 129 | value >>= 16; |
| 130 | bit += 16; |
| 131 | } |
| 132 | |
| 133 | if (value >= 1 << 8) { |
| 134 | value >>= 8; |
| 135 | bit += 8; |
| 136 | } |
| 137 | |
| 138 | if (value >= 1 << 4) { |
| 139 | value >>= 4; |
| 140 | bit += 4; |
| 141 | } |
| 142 | |
| 143 | if (value >= 1 << 2) { |
| 144 | value >>= 2; |
| 145 | bit += 2; |
| 146 | } |
| 147 | |
| 148 | if (value >= 1 << 1) { |
| 149 | //value >>= 1; |
| 150 | bit += 1; |
| 151 | } |
| 152 | |
| 153 | return bit; |
| 154 | } |
| 155 | |
| 156 | static int32_t |
| 157 | getPairIndex(UChar32 ch) |
| 158 | { |
| 159 | int32_t pairedCharCount = UPRV_LENGTHOF(pairedChars); |
| 160 | int32_t pairedCharPower = 1 << highBit(pairedCharCount); |
| 161 | int32_t = pairedCharCount - pairedCharPower; |
| 162 | |
| 163 | int32_t probe = pairedCharPower; |
| 164 | int32_t pairIndex = 0; |
| 165 | |
| 166 | if (ch >= pairedChars[pairedCharExtra]) { |
| 167 | pairIndex = pairedCharExtra; |
| 168 | } |
| 169 | |
| 170 | while (probe > (1 << 0)) { |
| 171 | probe >>= 1; |
| 172 | |
| 173 | if (ch >= pairedChars[pairIndex + probe]) { |
| 174 | pairIndex += probe; |
| 175 | } |
| 176 | } |
| 177 | |
| 178 | if (pairedChars[pairIndex] != ch) { |
| 179 | pairIndex = -1; |
| 180 | } |
| 181 | |
| 182 | return pairIndex; |
| 183 | } |
| 184 | |
| 185 | static UBool |
| 186 | sameScript(UScriptCode scriptOne, UScriptCode scriptTwo) |
| 187 | { |
| 188 | return scriptOne <= USCRIPT_INHERITED || scriptTwo <= USCRIPT_INHERITED || scriptOne == scriptTwo; |
| 189 | } |
| 190 | |
| 191 | U_CAPI UScriptRun * U_EXPORT2 |
| 192 | uscript_openRun(const UChar *src, int32_t length, UErrorCode *pErrorCode) |
| 193 | { |
| 194 | UScriptRun *result = NULL; |
| 195 | |
| 196 | if (pErrorCode == NULL || U_FAILURE(*pErrorCode)) { |
| 197 | return NULL; |
| 198 | } |
| 199 | |
| 200 | result = (UScriptRun *)uprv_malloc(sizeof (UScriptRun)); |
| 201 | |
| 202 | if (result == NULL) { |
| 203 | *pErrorCode = U_MEMORY_ALLOCATION_ERROR; |
| 204 | return NULL; |
| 205 | } |
| 206 | |
| 207 | uscript_setRunText(result, src, length, pErrorCode); |
| 208 | |
| 209 | /* Release the UScriptRun if uscript_setRunText() returns an error */ |
| 210 | if (U_FAILURE(*pErrorCode)) { |
| 211 | uprv_free(result); |
| 212 | result = NULL; |
| 213 | } |
| 214 | |
| 215 | return result; |
| 216 | } |
| 217 | |
| 218 | U_CAPI void U_EXPORT2 |
| 219 | uscript_closeRun(UScriptRun *scriptRun) |
| 220 | { |
| 221 | if (scriptRun != NULL) { |
| 222 | uprv_free(scriptRun); |
| 223 | } |
| 224 | } |
| 225 | |
| 226 | U_CAPI void U_EXPORT2 |
| 227 | uscript_resetRun(UScriptRun *scriptRun) |
| 228 | { |
| 229 | if (scriptRun != NULL) { |
| 230 | scriptRun->scriptStart = 0; |
| 231 | scriptRun->scriptLimit = 0; |
| 232 | scriptRun->scriptCode = USCRIPT_INVALID_CODE; |
| 233 | scriptRun->parenSP = -1; |
| 234 | scriptRun->pushCount = 0; |
| 235 | scriptRun->fixupCount = 0; |
| 236 | } |
| 237 | } |
| 238 | |
| 239 | U_CAPI void U_EXPORT2 |
| 240 | uscript_setRunText(UScriptRun *scriptRun, const UChar *src, int32_t length, UErrorCode *pErrorCode) |
| 241 | { |
| 242 | if (pErrorCode == NULL || U_FAILURE(*pErrorCode)) { |
| 243 | return; |
| 244 | } |
| 245 | |
| 246 | if (scriptRun == NULL || length < 0 || ((src == NULL) != (length == 0))) { |
| 247 | *pErrorCode = U_ILLEGAL_ARGUMENT_ERROR; |
| 248 | return; |
| 249 | } |
| 250 | |
| 251 | scriptRun->textArray = src; |
| 252 | scriptRun->textLength = length; |
| 253 | |
| 254 | uscript_resetRun(scriptRun); |
| 255 | } |
| 256 | |
| 257 | U_CAPI UBool U_EXPORT2 |
| 258 | uscript_nextRun(UScriptRun *scriptRun, int32_t *pRunStart, int32_t *pRunLimit, UScriptCode *pRunScript) |
| 259 | { |
| 260 | UErrorCode error = U_ZERO_ERROR; |
| 261 | |
| 262 | /* if we've fallen off the end of the text, we're done */ |
| 263 | if (scriptRun == NULL || scriptRun->scriptLimit >= scriptRun->textLength) { |
| 264 | return FALSE; |
| 265 | } |
| 266 | |
| 267 | SYNC_FIXUP(scriptRun); |
| 268 | scriptRun->scriptCode = USCRIPT_COMMON; |
| 269 | |
| 270 | for (scriptRun->scriptStart = scriptRun->scriptLimit; scriptRun->scriptLimit < scriptRun->textLength; scriptRun->scriptLimit += 1) { |
| 271 | UChar high = scriptRun->textArray[scriptRun->scriptLimit]; |
| 272 | UChar32 ch = high; |
| 273 | UScriptCode sc; |
| 274 | int32_t pairIndex; |
| 275 | |
| 276 | /* |
| 277 | * if the character is a high surrogate and it's not the last one |
| 278 | * in the text, see if it's followed by a low surrogate |
| 279 | */ |
| 280 | if (high >= 0xD800 && high <= 0xDBFF && scriptRun->scriptLimit < scriptRun->textLength - 1) { |
| 281 | UChar low = scriptRun->textArray[scriptRun->scriptLimit + 1]; |
| 282 | |
| 283 | /* |
| 284 | * if it is followed by a low surrogate, |
| 285 | * consume it and form the full character |
| 286 | */ |
| 287 | if (low >= 0xDC00 && low <= 0xDFFF) { |
| 288 | ch = (high - 0xD800) * 0x0400 + low - 0xDC00 + 0x10000; |
| 289 | scriptRun->scriptLimit += 1; |
| 290 | } |
| 291 | } |
| 292 | |
| 293 | sc = uscript_getScript(ch, &error); |
| 294 | pairIndex = getPairIndex(ch); |
| 295 | |
| 296 | /* |
| 297 | * Paired character handling: |
| 298 | * |
| 299 | * if it's an open character, push it onto the stack. |
| 300 | * if it's a close character, find the matching open on the |
| 301 | * stack, and use that script code. Any non-matching open |
| 302 | * characters above it on the stack will be poped. |
| 303 | */ |
| 304 | if (pairIndex >= 0) { |
| 305 | if ((pairIndex & 1) == 0) { |
| 306 | push(scriptRun, pairIndex, scriptRun->scriptCode); |
| 307 | } else { |
| 308 | int32_t pi = pairIndex & ~1; |
| 309 | |
| 310 | while (STACK_IS_NOT_EMPTY(scriptRun) && TOP(scriptRun).pairIndex != pi) { |
| 311 | pop(scriptRun); |
| 312 | } |
| 313 | |
| 314 | if (STACK_IS_NOT_EMPTY(scriptRun)) { |
| 315 | sc = TOP(scriptRun).scriptCode; |
| 316 | } |
| 317 | } |
| 318 | } |
| 319 | |
| 320 | if (sameScript(scriptRun->scriptCode, sc)) { |
| 321 | if (scriptRun->scriptCode <= USCRIPT_INHERITED && sc > USCRIPT_INHERITED) { |
| 322 | scriptRun->scriptCode = sc; |
| 323 | |
| 324 | fixup(scriptRun, scriptRun->scriptCode); |
| 325 | } |
| 326 | |
| 327 | /* |
| 328 | * if this character is a close paired character, |
| 329 | * pop the matching open character from the stack |
| 330 | */ |
| 331 | if (pairIndex >= 0 && (pairIndex & 1) != 0) { |
| 332 | pop(scriptRun); |
| 333 | } |
| 334 | } else { |
| 335 | /* |
| 336 | * if the run broke on a surrogate pair, |
| 337 | * end it before the high surrogate |
| 338 | */ |
| 339 | if (ch >= 0x10000) { |
| 340 | scriptRun->scriptLimit -= 1; |
| 341 | } |
| 342 | |
| 343 | break; |
| 344 | } |
| 345 | } |
| 346 | |
| 347 | |
| 348 | if (pRunStart != NULL) { |
| 349 | *pRunStart = scriptRun->scriptStart; |
| 350 | } |
| 351 | |
| 352 | if (pRunLimit != NULL) { |
| 353 | *pRunLimit = scriptRun->scriptLimit; |
| 354 | } |
| 355 | |
| 356 | if (pRunScript != NULL) { |
| 357 | *pRunScript = scriptRun->scriptCode; |
| 358 | } |
| 359 | |
| 360 | return TRUE; |
| 361 | } |
| 362 | |