| 1 |
/* Case mapping for UTF-8/UTF-16/UTF-32 strings (locale dependent). |
| 2 |
Copyright (C) 2009-2024 Free Software Foundation, Inc. |
| 3 |
Written by Bruno Haible <[email protected]>, 2009. |
| 4 |
|
| 5 |
This file is free software. |
| 6 |
It is dual-licensed under "the GNU LGPLv3+ or the GNU GPLv2+". |
| 7 |
You can redistribute it and/or modify it under either |
| 8 |
- the terms of the GNU Lesser General Public License as published |
| 9 |
by the Free Software Foundation, either version 3, or (at your |
| 10 |
option) any later version, or |
| 11 |
- the terms of the GNU General Public License as published by the |
| 12 |
Free Software Foundation; either version 2, or (at your option) |
| 13 |
any later version, or |
| 14 |
- the same dual license "the GNU LGPLv3+ or the GNU GPLv2+". |
| 15 |
|
| 16 |
This file is distributed in the hope that it will be useful, |
| 17 |
but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 18 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 19 |
Lesser General Public License and the GNU General Public License |
| 20 |
for more details. |
| 21 |
|
| 22 |
You should have received a copy of the GNU Lesser General Public |
| 23 |
License and of the GNU General Public License along with this |
| 24 |
program. If not, see <https://www.gnu.org/licenses/>. */ |
| 25 |
|
| 26 |
UNIT * |
| 27 |
FUNC (const UNIT *s, size_t n, |
| 28 |
casing_prefix_context_t prefix_context, |
| 29 |
casing_suffix_context_t suffix_context, |
| 30 |
const char *iso639_language, |
| 31 |
ucs4_t (*single_character_map) (ucs4_t), |
| 32 |
size_t offset_in_rule, /* offset in 'struct special_casing_rule' */ |
| 33 |
uninorm_t nf, |
| 34 |
UNIT *resultbuf, size_t *lengthp) |
| 35 |
{ |
| 36 |
/* The result being accumulated. */ |
| 37 |
UNIT *result; |
| 38 |
size_t length; |
| 39 |
size_t allocated; |
| 40 |
|
| 41 |
/* Initialize the accumulator. */ |
| 42 |
if (nf != NULL || resultbuf == NULL) |
| 43 |
{ |
| 44 |
result = NULL; |
| 45 |
allocated = 0; |
| 46 |
} |
| 47 |
else |
| 48 |
{ |
| 49 |
result = resultbuf; |
| 50 |
allocated = *lengthp; |
| 51 |
} |
| 52 |
length = 0; |
| 53 |
|
| 54 |
{ |
| 55 |
const UNIT *s_end = s + n; |
| 56 |
|
| 57 |
/* Helper for evaluating the FINAL_SIGMA condition: |
| 58 |
Last character that was not case-ignorable. */ |
| 59 |
ucs4_t last_char_except_ignorable = |
| 60 |
prefix_context.last_char_except_ignorable; |
| 61 |
|
| 62 |
/* Helper for evaluating the AFTER_SOFT_DOTTED and AFTER_I conditions: |
| 63 |
Last character that was of combining class 230 ("Above") or 0. */ |
| 64 |
ucs4_t last_char_normal_or_above = |
| 65 |
prefix_context.last_char_normal_or_above; |
| 66 |
|
| 67 |
while (s < s_end) |
| 68 |
{ |
| 69 |
ucs4_t uc; |
| 70 |
int count = U_MBTOUC_UNSAFE (&uc, s, s_end - s); |
| 71 |
|
| 72 |
ucs4_t mapped_uc[3]; |
| 73 |
unsigned int mapped_count; |
| 74 |
|
| 75 |
if (uc < 0x10000) |
| 76 |
{ |
| 77 |
/* Look first in the special-casing table. */ |
| 78 |
char code[3]; |
| 79 |
|
| 80 |
code[0] = (uc >> 8) & 0xff; |
| 81 |
code[1] = uc & 0xff; |
| 82 |
|
| 83 |
for (code[2] = 0; ; code[2]++) |
| 84 |
{ |
| 85 |
const struct special_casing_rule *rule = |
| 86 |
gl_unicase_special_lookup (code, 3); |
| 87 |
|
| 88 |
if (rule == NULL) |
| 89 |
break; |
| 90 |
|
| 91 |
/* Test if the condition applies. */ |
| 92 |
/* Does the language apply? */ |
| 93 |
if (rule->language[0] == '\0' |
| 94 |
|| (iso639_language != NULL |
| 95 |
&& iso639_language[0] == rule->language[0] |
| 96 |
&& iso639_language[1] == rule->language[1])) |
| 97 |
{ |
| 98 |
/* Does the context apply? */ |
| 99 |
int context = rule->context; |
| 100 |
bool applies; |
| 101 |
|
| 102 |
if (context < 0) |
| 103 |
context = - context; |
| 104 |
switch (context) |
| 105 |
{ |
| 106 |
case SCC_ALWAYS: |
| 107 |
applies = true; |
| 108 |
break; |
| 109 |
|
| 110 |
case SCC_FINAL_SIGMA: |
| 111 |
/* "Before" condition: preceded by a sequence |
| 112 |
consisting of a cased letter and a case-ignorable |
| 113 |
sequence. |
| 114 |
"After" condition: not followed by a sequence |
| 115 |
consisting of a case-ignorable sequence and then a |
| 116 |
cased letter. */ |
| 117 |
/* Test the "before" condition. */ |
| 118 |
applies = uc_is_cased (last_char_except_ignorable); |
| 119 |
/* Test the "after" condition. */ |
| 120 |
if (applies) |
| 121 |
{ |
| 122 |
const UNIT *s2 = s + count; |
| 123 |
for (;;) |
| 124 |
{ |
| 125 |
if (s2 < s_end) |
| 126 |
{ |
| 127 |
ucs4_t uc2; |
| 128 |
int count2 = U_MBTOUC_UNSAFE (&uc2, s2, s_end - s2); |
| 129 |
/* Our uc_is_case_ignorable function is |
| 130 |
known to return false for all cased |
| 131 |
characters. So we can call |
| 132 |
uc_is_case_ignorable first. */ |
| 133 |
if (!uc_is_case_ignorable (uc2)) |
| 134 |
{ |
| 135 |
applies = ! uc_is_cased (uc2); |
| 136 |
break; |
| 137 |
} |
| 138 |
s2 += count2; |
| 139 |
} |
| 140 |
else |
| 141 |
{ |
| 142 |
applies = ! uc_is_cased (suffix_context.first_char_except_ignorable); |
| 143 |
break; |
| 144 |
} |
| 145 |
} |
| 146 |
} |
| 147 |
break; |
| 148 |
|
| 149 |
case SCC_AFTER_SOFT_DOTTED: |
| 150 |
/* "Before" condition: There is a Soft_Dotted character |
| 151 |
before it, with no intervening character of |
| 152 |
combining class 0 or 230 (Above). */ |
| 153 |
/* Test the "before" condition. */ |
| 154 |
applies = uc_is_property_soft_dotted (last_char_normal_or_above); |
| 155 |
break; |
| 156 |
|
| 157 |
case SCC_MORE_ABOVE: |
| 158 |
/* "After" condition: followed by a character of |
| 159 |
combining class 230 (Above) with no intervening |
| 160 |
character of combining class 0 or 230 (Above). */ |
| 161 |
/* Test the "after" condition. */ |
| 162 |
{ |
| 163 |
const UNIT *s2 = s + count; |
| 164 |
applies = false; |
| 165 |
for (;;) |
| 166 |
{ |
| 167 |
if (s2 < s_end) |
| 168 |
{ |
| 169 |
ucs4_t uc2; |
| 170 |
int count2 = U_MBTOUC_UNSAFE (&uc2, s2, s_end - s2); |
| 171 |
int ccc = uc_combining_class (uc2); |
| 172 |
if (ccc == UC_CCC_A) |
| 173 |
{ |
| 174 |
applies = true; |
| 175 |
break; |
| 176 |
} |
| 177 |
if (ccc == UC_CCC_NR) |
| 178 |
break; |
| 179 |
s2 += count2; |
| 180 |
} |
| 181 |
else |
| 182 |
{ |
| 183 |
applies = ((suffix_context.bits & SCC_MORE_ABOVE_MASK) != 0); |
| 184 |
break; |
| 185 |
} |
| 186 |
} |
| 187 |
} |
| 188 |
break; |
| 189 |
|
| 190 |
case SCC_BEFORE_DOT: |
| 191 |
/* "After" condition: followed by COMBINING DOT ABOVE |
| 192 |
(U+0307). Any sequence of characters with a |
| 193 |
combining class that is neither 0 nor 230 may |
| 194 |
intervene between the current character and the |
| 195 |
combining dot above. */ |
| 196 |
/* Test the "after" condition. */ |
| 197 |
{ |
| 198 |
const UNIT *s2 = s + count; |
| 199 |
applies = false; |
| 200 |
for (;;) |
| 201 |
{ |
| 202 |
if (s2 < s_end) |
| 203 |
{ |
| 204 |
ucs4_t uc2; |
| 205 |
int count2 = U_MBTOUC_UNSAFE (&uc2, s2, s_end - s2); |
| 206 |
if (uc2 == 0x0307) /* COMBINING DOT ABOVE */ |
| 207 |
{ |
| 208 |
applies = true; |
| 209 |
break; |
| 210 |
} |
| 211 |
{ |
| 212 |
int ccc = uc_combining_class (uc2); |
| 213 |
if (ccc == UC_CCC_A || ccc == UC_CCC_NR) |
| 214 |
break; |
| 215 |
} |
| 216 |
s2 += count2; |
| 217 |
} |
| 218 |
else |
| 219 |
{ |
| 220 |
applies = ((suffix_context.bits & SCC_BEFORE_DOT_MASK) != 0); |
| 221 |
break; |
| 222 |
} |
| 223 |
} |
| 224 |
} |
| 225 |
break; |
| 226 |
|
| 227 |
case SCC_AFTER_I: |
| 228 |
/* "Before" condition: There is an uppercase I before |
| 229 |
it, and there is no intervening character of |
| 230 |
combining class 0 or 230 (Above). */ |
| 231 |
/* Test the "before" condition. */ |
| 232 |
applies = (last_char_normal_or_above == 'I'); |
| 233 |
break; |
| 234 |
|
| 235 |
default: |
| 236 |
abort (); |
| 237 |
} |
| 238 |
if (rule->context < 0) |
| 239 |
applies = !applies; |
| 240 |
|
| 241 |
if (applies) |
| 242 |
{ |
| 243 |
/* The rule applies. |
| 244 |
Look up the mapping (0 to 3 characters). */ |
| 245 |
const unsigned short *mapped_in_rule = |
| 246 |
(const unsigned short *)((const char *)rule + offset_in_rule); |
| 247 |
|
| 248 |
if (mapped_in_rule[0] == 0) |
| 249 |
mapped_count = 0; |
| 250 |
else |
| 251 |
{ |
| 252 |
mapped_uc[0] = mapped_in_rule[0]; |
| 253 |
if (mapped_in_rule[1] == 0) |
| 254 |
mapped_count = 1; |
| 255 |
else |
| 256 |
{ |
| 257 |
mapped_uc[1] = mapped_in_rule[1]; |
| 258 |
if (mapped_in_rule[2] == 0) |
| 259 |
mapped_count = 2; |
| 260 |
else |
| 261 |
{ |
| 262 |
mapped_uc[2] = mapped_in_rule[2]; |
| 263 |
mapped_count = 3; |
| 264 |
} |
| 265 |
} |
| 266 |
} |
| 267 |
goto found_mapping; |
| 268 |
} |
| 269 |
} |
| 270 |
|
| 271 |
/* Optimization: Save a hash table lookup in the next round. */ |
| 272 |
if (!rule->has_next) |
| 273 |
break; |
| 274 |
} |
| 275 |
} |
| 276 |
|
| 277 |
/* No special-cased mapping. So use the locale and context independent |
| 278 |
mapping. */ |
| 279 |
mapped_uc[0] = single_character_map (uc); |
| 280 |
mapped_count = 1; |
| 281 |
|
| 282 |
found_mapping: |
| 283 |
/* Found the mapping: uc maps to mapped_uc[0..mapped_count-1]. */ |
| 284 |
{ |
| 285 |
unsigned int i; |
| 286 |
|
| 287 |
for (i = 0; i < mapped_count; i++) |
| 288 |
{ |
| 289 |
ucs4_t muc = mapped_uc[i]; |
| 290 |
|
| 291 |
/* Append muc to the result accumulator. */ |
| 292 |
if (length < allocated) |
| 293 |
{ |
| 294 |
int ret = U_UCTOMB (result + length, muc, allocated - length); |
| 295 |
if (ret == -1) |
| 296 |
{ |
| 297 |
errno = EINVAL; |
| 298 |
goto fail; |
| 299 |
} |
| 300 |
if (ret >= 0) |
| 301 |
{ |
| 302 |
length += ret; |
| 303 |
goto done_appending; |
| 304 |
} |
| 305 |
} |
| 306 |
{ |
| 307 |
size_t old_allocated = allocated; |
| 308 |
size_t new_allocated = 2 * old_allocated; |
| 309 |
if (new_allocated < 64) |
| 310 |
new_allocated = 64; |
| 311 |
if (new_allocated < old_allocated) /* integer overflow? */ |
| 312 |
abort (); |
| 313 |
{ |
| 314 |
UNIT *larger_result; |
| 315 |
if (result == NULL) |
| 316 |
{ |
| 317 |
larger_result = (UNIT *) malloc (new_allocated * sizeof (UNIT)); |
| 318 |
if (larger_result == NULL) |
| 319 |
{ |
| 320 |
errno = ENOMEM; |
| 321 |
goto fail; |
| 322 |
} |
| 323 |
} |
| 324 |
else if (result == resultbuf) |
| 325 |
{ |
| 326 |
larger_result = (UNIT *) malloc (new_allocated * sizeof (UNIT)); |
| 327 |
if (larger_result == NULL) |
| 328 |
{ |
| 329 |
errno = ENOMEM; |
| 330 |
goto fail; |
| 331 |
} |
| 332 |
U_CPY (larger_result, resultbuf, length); |
| 333 |
} |
| 334 |
else |
| 335 |
{ |
| 336 |
larger_result = |
| 337 |
(UNIT *) realloc (result, new_allocated * sizeof (UNIT)); |
| 338 |
if (larger_result == NULL) |
| 339 |
{ |
| 340 |
errno = ENOMEM; |
| 341 |
goto fail; |
| 342 |
} |
| 343 |
} |
| 344 |
result = larger_result; |
| 345 |
allocated = new_allocated; |
| 346 |
{ |
| 347 |
int ret = U_UCTOMB (result + length, muc, allocated - length); |
| 348 |
if (ret == -1) |
| 349 |
{ |
| 350 |
errno = EINVAL; |
| 351 |
goto fail; |
| 352 |
} |
| 353 |
if (ret < 0) |
| 354 |
abort (); |
| 355 |
length += ret; |
| 356 |
goto done_appending; |
| 357 |
} |
| 358 |
} |
| 359 |
} |
| 360 |
done_appending: ; |
| 361 |
} |
| 362 |
} |
| 363 |
|
| 364 |
if (!uc_is_case_ignorable (uc)) |
| 365 |
last_char_except_ignorable = uc; |
| 366 |
|
| 367 |
{ |
| 368 |
int ccc = uc_combining_class (uc); |
| 369 |
if (ccc == UC_CCC_A || ccc == UC_CCC_NR) |
| 370 |
last_char_normal_or_above = uc; |
| 371 |
} |
| 372 |
|
| 373 |
s += count; |
| 374 |
} |
| 375 |
} |
| 376 |
|
| 377 |
if (nf != NULL) |
| 378 |
{ |
| 379 |
/* Finally, normalize the result. */ |
| 380 |
UNIT *normalized_result; |
| 381 |
|
| 382 |
normalized_result = U_NORMALIZE (nf, result, length, resultbuf, lengthp); |
| 383 |
if (normalized_result == NULL) |
| 384 |
goto fail; |
| 385 |
|
| 386 |
free (result); |
| 387 |
return normalized_result; |
| 388 |
} |
| 389 |
|
| 390 |
if (length == 0) |
| 391 |
{ |
| 392 |
if (result == NULL) |
| 393 |
{ |
| 394 |
/* Return a non-NULL value. NULL means error. */ |
| 395 |
result = (UNIT *) malloc (1); |
| 396 |
if (result == NULL) |
| 397 |
{ |
| 398 |
errno = ENOMEM; |
| 399 |
goto fail; |
| 400 |
} |
| 401 |
} |
| 402 |
} |
| 403 |
else if (result != resultbuf && length < allocated) |
| 404 |
{ |
| 405 |
/* Shrink the allocated memory if possible. */ |
| 406 |
UNIT *memory; |
| 407 |
|
| 408 |
memory = (UNIT *) realloc (result, length * sizeof (UNIT)); |
| 409 |
if (memory != NULL) |
| 410 |
result = memory; |
| 411 |
} |
| 412 |
|
| 413 |
*lengthp = length; |
| 414 |
return result; |
| 415 |
|
| 416 |
fail: |
| 417 |
if (result != resultbuf) |
| 418 |
{ |
| 419 |
int saved_errno = errno; |
| 420 |
free (result); |
| 421 |
errno = saved_errno; |
| 422 |
} |
| 423 |
return NULL; |
| 424 |
} |