1 /* utf8.c 2 * 3 * Copyright (c) 1998-2002, Larry Wall 4 * 5 * You may distribute under the terms of either the GNU General Public 6 * License or the Artistic License, as specified in the README file. 7 * 8 */ 9 10 /* 11 * 'What a fix!' said Sam. 'That's the one place in all the lands we've ever 12 * heard of that we don't want to see any closer; and that's the one place 13 * we're trying to get to! And that's just where we can't get, nohow.' 14 * 15 * 'Well do I understand your speech,' he answered in the same language; 16 * 'yet few strangers do so. Why then do you not speak in the Common Tongue, 17 * as is the custom in the West, if you wish to be answered?' 18 * 19 * ...the travellers perceived that the floor was paved with stones of many 20 * hues; branching runes and strange devices intertwined beneath their feet. 21 */ 22 23 #include "EXTERN.h" 24 #define PERL_IN_UTF8_C 25 #include "perl.h" 26 27 static char unees[] = "Malformed UTF-8 character (unexpected end of string)"; 28 29 /* 30 =head1 Unicode Support 31 32 =for apidoc A|U8 *|uvuni_to_utf8_flags|U8 *d|UV uv|UV flags 33 34 Adds the UTF8 representation of the Unicode codepoint C<uv> to the end 35 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free 36 bytes available. The return value is the pointer to the byte after the 37 end of the new character. In other words, 38 39 d = uvuni_to_utf8_flags(d, uv, flags); 40 41 or, in most cases, 42 43 d = uvuni_to_utf8(d, uv); 44 45 (which is equivalent to) 46 47 d = uvuni_to_utf8_flags(d, uv, 0); 48 49 is the recommended Unicode-aware way of saying 50 51 *(d++) = uv; 52 53 =cut 54 */ 55 56 U8 * 57 Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) 58 { 59 if (ckWARN(WARN_UTF8)) { 60 if (UNICODE_IS_SURROGATE(uv) && 61 !(flags & UNICODE_ALLOW_SURROGATE)) 62 Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv); 63 else if ( 64 ((uv >= 0xFDD0 && uv <= 0xFDEF && 65 !(flags & UNICODE_ALLOW_FDD0)) 66 || 67 ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */ 68 !(flags & UNICODE_ALLOW_FFFF))) && 69 /* UNICODE_ALLOW_SUPER includes 70 * FFFEs and FFFFs beyond 0x10FFFF. */ 71 ((uv <= PERL_UNICODE_MAX) || 72 !(flags & UNICODE_ALLOW_SUPER)) 73 ) 74 Perl_warner(aTHX_ packWARN(WARN_UTF8), 75 "Unicode character 0x%04"UVxf" is illegal", uv); 76 } 77 if (UNI_IS_INVARIANT(uv)) { 78 *d++ = (U8)UTF_TO_NATIVE(uv); 79 return d; 80 } 81 #if defined(EBCDIC) 82 else { 83 STRLEN len = UNISKIP(uv); 84 U8 *p = d+len-1; 85 while (p > d) { 86 *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK); 87 uv >>= UTF_ACCUMULATION_SHIFT; 88 } 89 *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len)); 90 return d+len; 91 } 92 #else /* Non loop style */ 93 if (uv < 0x800) { 94 *d++ = (U8)(( uv >> 6) | 0xc0); 95 *d++ = (U8)(( uv & 0x3f) | 0x80); 96 return d; 97 } 98 if (uv < 0x10000) { 99 *d++ = (U8)(( uv >> 12) | 0xe0); 100 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); 101 *d++ = (U8)(( uv & 0x3f) | 0x80); 102 return d; 103 } 104 if (uv < 0x200000) { 105 *d++ = (U8)(( uv >> 18) | 0xf0); 106 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); 107 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); 108 *d++ = (U8)(( uv & 0x3f) | 0x80); 109 return d; 110 } 111 if (uv < 0x4000000) { 112 *d++ = (U8)(( uv >> 24) | 0xf8); 113 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); 114 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); 115 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); 116 *d++ = (U8)(( uv & 0x3f) | 0x80); 117 return d; 118 } 119 if (uv < 0x80000000) { 120 *d++ = (U8)(( uv >> 30) | 0xfc); 121 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); 122 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); 123 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); 124 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); 125 *d++ = (U8)(( uv & 0x3f) | 0x80); 126 return d; 127 } 128 #ifdef HAS_QUAD 129 if (uv < UTF8_QUAD_MAX) 130 #endif 131 { 132 *d++ = 0xfe; /* Can't match U+FEFF! */ 133 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); 134 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); 135 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); 136 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); 137 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); 138 *d++ = (U8)(( uv & 0x3f) | 0x80); 139 return d; 140 } 141 #ifdef HAS_QUAD 142 { 143 *d++ = 0xff; /* Can't match U+FFFE! */ 144 *d++ = 0x80; /* 6 Reserved bits */ 145 *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */ 146 *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80); 147 *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80); 148 *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80); 149 *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80); 150 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); 151 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); 152 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); 153 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); 154 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); 155 *d++ = (U8)(( uv & 0x3f) | 0x80); 156 return d; 157 } 158 #endif 159 #endif /* Loop style */ 160 } 161 162 U8 * 163 Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv) 164 { 165 return Perl_uvuni_to_utf8_flags(aTHX_ d, uv, 0); 166 } 167 168 169 /* 170 =for apidoc A|STRLEN|is_utf8_char|U8 *s 171 172 Tests if some arbitrary number of bytes begins in a valid UTF-8 173 character. Note that an INVARIANT (i.e. ASCII) character is a valid 174 UTF-8 character. The actual number of bytes in the UTF-8 character 175 will be returned if it is valid, otherwise 0. 176 177 =cut */ 178 STRLEN 179 Perl_is_utf8_char(pTHX_ U8 *s) 180 { 181 U8 u = *s; 182 STRLEN slen, len; 183 UV uv, ouv; 184 185 if (UTF8_IS_INVARIANT(u)) 186 return 1; 187 188 if (!UTF8_IS_START(u)) 189 return 0; 190 191 len = UTF8SKIP(s); 192 193 if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) 194 return 0; 195 196 slen = len - 1; 197 s++; 198 u &= UTF_START_MASK(len); 199 uv = u; 200 ouv = uv; 201 while (slen--) { 202 if (!UTF8_IS_CONTINUATION(*s)) 203 return 0; 204 uv = UTF8_ACCUMULATE(uv, *s); 205 if (uv < ouv) 206 return 0; 207 ouv = uv; 208 s++; 209 } 210 211 if ((STRLEN)UNISKIP(uv) < len) 212 return 0; 213 214 return len; 215 } 216 217 /* 218 =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len 219 220 Returns true if first C<len> bytes of the given string form a valid UTF8 221 string, false otherwise. Note that 'a valid UTF8 string' does not mean 222 'a string that contains UTF8' because a valid ASCII string is a valid 223 UTF8 string. 224 225 =cut 226 */ 227 228 bool 229 Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len) 230 { 231 U8* x = s; 232 U8* send; 233 STRLEN c; 234 235 if (!len) 236 len = strlen((char *)s); 237 send = s + len; 238 239 while (x < send) { 240 /* Inline the easy bits of is_utf8_char() here for speed... */ 241 if (UTF8_IS_INVARIANT(*x)) 242 c = 1; 243 else if (!UTF8_IS_START(*x)) 244 return FALSE; 245 else { 246 /* ... and call is_utf8_char() only if really needed. */ 247 c = is_utf8_char(x); 248 if (!c) 249 return FALSE; 250 } 251 x += c; 252 } 253 if (x != send) 254 return FALSE; 255 256 return TRUE; 257 } 258 259 /* 260 =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags 261 262 Bottom level UTF-8 decode routine. 263 Returns the unicode code point value of the first character in the string C<s> 264 which is assumed to be in UTF8 encoding and no longer than C<curlen>; 265 C<retlen> will be set to the length, in bytes, of that character. 266 267 If C<s> does not point to a well-formed UTF8 character, the behaviour 268 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY, 269 it is assumed that the caller will raise a warning, and this function 270 will silently just set C<retlen> to C<-1> and return zero. If the 271 C<flags> does not contain UTF8_CHECK_ONLY, warnings about 272 malformations will be given, C<retlen> will be set to the expected 273 length of the UTF-8 character in bytes, and zero will be returned. 274 275 The C<flags> can also contain various flags to allow deviations from 276 the strict UTF-8 encoding (see F<utf8.h>). 277 278 Most code should use utf8_to_uvchr() rather than call this directly. 279 280 =cut 281 */ 282 283 UV 284 Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) 285 { 286 U8 *s0 = s; 287 UV uv = *s, ouv = 0; 288 STRLEN len = 1; 289 bool dowarn = ckWARN_d(WARN_UTF8); 290 UV startbyte = *s; 291 STRLEN expectlen = 0; 292 U32 warning = 0; 293 294 /* This list is a superset of the UTF8_ALLOW_XXX. */ 295 296 #define UTF8_WARN_EMPTY 1 297 #define UTF8_WARN_CONTINUATION 2 298 #define UTF8_WARN_NON_CONTINUATION 3 299 #define UTF8_WARN_FE_FF 4 300 #define UTF8_WARN_SHORT 5 301 #define UTF8_WARN_OVERFLOW 6 302 #define UTF8_WARN_SURROGATE 7 303 #define UTF8_WARN_LONG 8 304 #define UTF8_WARN_FFFF 9 /* Also FFFE. */ 305 306 if (curlen == 0 && 307 !(flags & UTF8_ALLOW_EMPTY)) { 308 warning = UTF8_WARN_EMPTY; 309 goto malformed; 310 } 311 312 if (UTF8_IS_INVARIANT(uv)) { 313 if (retlen) 314 *retlen = 1; 315 return (UV) (NATIVE_TO_UTF(*s)); 316 } 317 318 if (UTF8_IS_CONTINUATION(uv) && 319 !(flags & UTF8_ALLOW_CONTINUATION)) { 320 warning = UTF8_WARN_CONTINUATION; 321 goto malformed; 322 } 323 324 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && 325 !(flags & UTF8_ALLOW_NON_CONTINUATION)) { 326 warning = UTF8_WARN_NON_CONTINUATION; 327 goto malformed; 328 } 329 330 #ifdef EBCDIC 331 uv = NATIVE_TO_UTF(uv); 332 #else 333 if ((uv == 0xfe || uv == 0xff) && 334 !(flags & UTF8_ALLOW_FE_FF)) { 335 warning = UTF8_WARN_FE_FF; 336 goto malformed; 337 } 338 #endif 339 340 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; } 341 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; } 342 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } 343 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } 344 #ifdef EBCDIC 345 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } 346 else { len = 7; uv &= 0x01; } 347 #else 348 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } 349 else if (!(uv & 0x01)) { len = 7; uv = 0; } 350 else { len = 13; uv = 0; } /* whoa! */ 351 #endif 352 353 if (retlen) 354 *retlen = len; 355 356 expectlen = len; 357 358 if ((curlen < expectlen) && 359 !(flags & UTF8_ALLOW_SHORT)) { 360 warning = UTF8_WARN_SHORT; 361 goto malformed; 362 } 363 364 len--; 365 s++; 366 ouv = uv; 367 368 while (len--) { 369 if (!UTF8_IS_CONTINUATION(*s) && 370 !(flags & UTF8_ALLOW_NON_CONTINUATION)) { 371 s--; 372 warning = UTF8_WARN_NON_CONTINUATION; 373 goto malformed; 374 } 375 else 376 uv = UTF8_ACCUMULATE(uv, *s); 377 if (!(uv > ouv)) { 378 /* These cannot be allowed. */ 379 if (uv == ouv) { 380 if (!(flags & UTF8_ALLOW_LONG)) { 381 warning = UTF8_WARN_LONG; 382 goto malformed; 383 } 384 } 385 else { /* uv < ouv */ 386 /* This cannot be allowed. */ 387 warning = UTF8_WARN_OVERFLOW; 388 goto malformed; 389 } 390 } 391 s++; 392 ouv = uv; 393 } 394 395 if (UNICODE_IS_SURROGATE(uv) && 396 !(flags & UTF8_ALLOW_SURROGATE)) { 397 warning = UTF8_WARN_SURROGATE; 398 goto malformed; 399 } else if ((expectlen > (STRLEN)UNISKIP(uv)) && 400 !(flags & UTF8_ALLOW_LONG)) { 401 warning = UTF8_WARN_LONG; 402 goto malformed; 403 } else if (UNICODE_IS_ILLEGAL(uv) && 404 !(flags & UTF8_ALLOW_FFFF)) { 405 warning = UTF8_WARN_FFFF; 406 goto malformed; 407 } 408 409 return uv; 410 411 malformed: 412 413 if (flags & UTF8_CHECK_ONLY) { 414 if (retlen) 415 *retlen = -1; 416 return 0; 417 } 418 419 if (dowarn) { 420 SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0)); 421 422 switch (warning) { 423 case 0: /* Intentionally empty. */ break; 424 case UTF8_WARN_EMPTY: 425 Perl_sv_catpvf(aTHX_ sv, "(empty string)"); 426 break; 427 case UTF8_WARN_CONTINUATION: 428 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv); 429 break; 430 case UTF8_WARN_NON_CONTINUATION: 431 if (s == s0) 432 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")", 433 (UV)s[1], startbyte); 434 else 435 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)", 436 (UV)s[1], s - s0, s - s0 > 1 ? "s" : "", startbyte, expectlen); 437 438 break; 439 case UTF8_WARN_FE_FF: 440 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv); 441 break; 442 case UTF8_WARN_SHORT: 443 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", 444 curlen, curlen == 1 ? "" : "s", expectlen, startbyte); 445 expectlen = curlen; /* distance for caller to skip */ 446 break; 447 case UTF8_WARN_OVERFLOW: 448 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")", 449 ouv, *s, startbyte); 450 break; 451 case UTF8_WARN_SURROGATE: 452 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv); 453 break; 454 case UTF8_WARN_LONG: 455 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", 456 expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte); 457 break; 458 case UTF8_WARN_FFFF: 459 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv); 460 break; 461 default: 462 Perl_sv_catpvf(aTHX_ sv, "(unknown reason)"); 463 break; 464 } 465 466 if (warning) { 467 char *s = SvPVX(sv); 468 469 if (PL_op) 470 Perl_warner(aTHX_ packWARN(WARN_UTF8), 471 "%s in %s", s, OP_DESC(PL_op)); 472 else 473 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s); 474 } 475 } 476 477 if (retlen) 478 *retlen = expectlen ? expectlen : len; 479 480 return 0; 481 } 482 483 /* 484 =for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen 485 486 Returns the native character value of the first character in the string C<s> 487 which is assumed to be in UTF8 encoding; C<retlen> will be set to the 488 length, in bytes, of that character. 489 490 If C<s> does not point to a well-formed UTF8 character, zero is 491 returned and retlen is set, if possible, to -1. 492 493 =cut 494 */ 495 496 UV 497 Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen) 498 { 499 return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen, 500 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); 501 } 502 503 /* 504 =for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen 505 506 Returns the Unicode code point of the first character in the string C<s> 507 which is assumed to be in UTF8 encoding; C<retlen> will be set to the 508 length, in bytes, of that character. 509 510 This function should only be used when returned UV is considered 511 an index into the Unicode semantic tables (e.g. swashes). 512 513 If C<s> does not point to a well-formed UTF8 character, zero is 514 returned and retlen is set, if possible, to -1. 515 516 =cut 517 */ 518 519 UV 520 Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen) 521 { 522 /* Call the low level routine asking for checks */ 523 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen, 524 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); 525 } 526 527 /* 528 =for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e 529 530 Return the length of the UTF-8 char encoded string C<s> in characters. 531 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end 532 up past C<e>, croaks. 533 534 =cut 535 */ 536 537 STRLEN 538 Perl_utf8_length(pTHX_ U8 *s, U8 *e) 539 { 540 STRLEN len = 0; 541 542 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. 543 * the bitops (especially ~) can create illegal UTF-8. 544 * In other words: in Perl UTF-8 is not just for Unicode. */ 545 546 if (e < s) { 547 if (ckWARN_d(WARN_UTF8)) { 548 if (PL_op) 549 Perl_warner(aTHX_ packWARN(WARN_UTF8), 550 "%s in %s", unees, OP_DESC(PL_op)); 551 else 552 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees); 553 } 554 return 0; 555 } 556 while (s < e) { 557 U8 t = UTF8SKIP(s); 558 559 if (e - s < t) { 560 if (ckWARN_d(WARN_UTF8)) { 561 if (PL_op) 562 Perl_warner(aTHX_ packWARN(WARN_UTF8), 563 unees, OP_DESC(PL_op)); 564 else 565 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees); 566 } 567 return len; 568 } 569 s += t; 570 len++; 571 } 572 573 return len; 574 } 575 576 /* 577 =for apidoc A|IV|utf8_distance|U8 *a|U8 *b 578 579 Returns the number of UTF8 characters between the UTF-8 pointers C<a> 580 and C<b>. 581 582 WARNING: use only if you *know* that the pointers point inside the 583 same UTF-8 buffer. 584 585 =cut 586 */ 587 588 IV 589 Perl_utf8_distance(pTHX_ U8 *a, U8 *b) 590 { 591 IV off = 0; 592 593 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. 594 * the bitops (especially ~) can create illegal UTF-8. 595 * In other words: in Perl UTF-8 is not just for Unicode. */ 596 597 if (a < b) { 598 while (a < b) { 599 U8 c = UTF8SKIP(a); 600 601 if (b - a < c) { 602 if (ckWARN_d(WARN_UTF8)) { 603 if (PL_op) 604 Perl_warner(aTHX_ packWARN(WARN_UTF8), 605 "%s in %s", unees, OP_DESC(PL_op)); 606 else 607 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees); 608 } 609 return off; 610 } 611 a += c; 612 off--; 613 } 614 } 615 else { 616 while (b < a) { 617 U8 c = UTF8SKIP(b); 618 619 if (a - b < c) { 620 if (ckWARN_d(WARN_UTF8)) { 621 if (PL_op) 622 Perl_warner(aTHX_ packWARN(WARN_UTF8), 623 "%s in %s", unees, OP_DESC(PL_op)); 624 else 625 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees); 626 } 627 return off; 628 } 629 b += c; 630 off++; 631 } 632 } 633 634 return off; 635 } 636 637 /* 638 =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off 639 640 Return the UTF-8 pointer C<s> displaced by C<off> characters, either 641 forward or backward. 642 643 WARNING: do not use the following unless you *know* C<off> is within 644 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned 645 on the first byte of character or just after the last byte of a character. 646 647 =cut 648 */ 649 650 U8 * 651 Perl_utf8_hop(pTHX_ U8 *s, I32 off) 652 { 653 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g 654 * the bitops (especially ~) can create illegal UTF-8. 655 * In other words: in Perl UTF-8 is not just for Unicode. */ 656 657 if (off >= 0) { 658 while (off--) 659 s += UTF8SKIP(s); 660 } 661 else { 662 while (off++) { 663 s--; 664 while (UTF8_IS_CONTINUATION(*s)) 665 s--; 666 } 667 } 668 return s; 669 } 670 671 /* 672 =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len 673 674 Converts a string C<s> of length C<len> from UTF8 into byte encoding. 675 Unlike C<bytes_to_utf8>, this over-writes the original string, and 676 updates len to contain the new length. 677 Returns zero on failure, setting C<len> to -1. 678 679 =cut 680 */ 681 682 U8 * 683 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) 684 { 685 U8 *send; 686 U8 *d; 687 U8 *save = s; 688 689 /* ensure valid UTF8 and chars < 256 before updating string */ 690 for (send = s + *len; s < send; ) { 691 U8 c = *s++; 692 693 if (!UTF8_IS_INVARIANT(c) && 694 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) 695 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { 696 *len = -1; 697 return 0; 698 } 699 } 700 701 d = s = save; 702 while (s < send) { 703 STRLEN ulen; 704 *d++ = (U8)utf8_to_uvchr(s, &ulen); 705 s += ulen; 706 } 707 *d = '\0'; 708 *len = d - save; 709 return save; 710 } 711 712 /* 713 =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8 714 715 Converts a string C<s> of length C<len> from UTF8 into byte encoding. 716 Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to 717 the newly-created string, and updates C<len> to contain the new 718 length. Returns the original string if no conversion occurs, C<len> 719 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to 720 0 if C<s> is converted or contains all 7bit characters. 721 722 =cut 723 */ 724 725 U8 * 726 Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8) 727 { 728 U8 *d; 729 U8 *start = s; 730 U8 *send; 731 I32 count = 0; 732 733 if (!*is_utf8) 734 return start; 735 736 /* ensure valid UTF8 and chars < 256 before converting string */ 737 for (send = s + *len; s < send;) { 738 U8 c = *s++; 739 if (!UTF8_IS_INVARIANT(c)) { 740 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send && 741 (c = *s++) && UTF8_IS_CONTINUATION(c)) 742 count++; 743 else 744 return start; 745 } 746 } 747 748 *is_utf8 = 0; 749 750 Newz(801, d, (*len) - count + 1, U8); 751 s = start; start = d; 752 while (s < send) { 753 U8 c = *s++; 754 if (!UTF8_IS_INVARIANT(c)) { 755 /* Then it is two-byte encoded */ 756 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++); 757 c = ASCII_TO_NATIVE(c); 758 } 759 *d++ = c; 760 } 761 *d = '\0'; 762 *len = d - start; 763 return start; 764 } 765 766 /* 767 =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len 768 769 Converts a string C<s> of length C<len> from ASCII into UTF8 encoding. 770 Returns a pointer to the newly-created string, and sets C<len> to 771 reflect the new length. 772 773 =cut 774 */ 775 776 U8* 777 Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len) 778 { 779 U8 *send; 780 U8 *d; 781 U8 *dst; 782 send = s + (*len); 783 784 Newz(801, d, (*len) * 2 + 1, U8); 785 dst = d; 786 787 while (s < send) { 788 UV uv = NATIVE_TO_ASCII(*s++); 789 if (UNI_IS_INVARIANT(uv)) 790 *d++ = (U8)UTF_TO_NATIVE(uv); 791 else { 792 *d++ = (U8)UTF8_EIGHT_BIT_HI(uv); 793 *d++ = (U8)UTF8_EIGHT_BIT_LO(uv); 794 } 795 } 796 *d = '\0'; 797 *len = d-dst; 798 return dst; 799 } 800 801 /* 802 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8. 803 * 804 * Destination must be pre-extended to 3/2 source. Do not use in-place. 805 * We optimize for native, for obvious reasons. */ 806 807 U8* 808 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) 809 { 810 U8* pend; 811 U8* dstart = d; 812 813 if (bytelen & 1) 814 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen"); 815 816 pend = p + bytelen; 817 818 while (p < pend) { 819 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ 820 p += 2; 821 if (uv < 0x80) { 822 *d++ = (U8)uv; 823 continue; 824 } 825 if (uv < 0x800) { 826 *d++ = (U8)(( uv >> 6) | 0xc0); 827 *d++ = (U8)(( uv & 0x3f) | 0x80); 828 continue; 829 } 830 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */ 831 UV low = *p++; 832 if (low < 0xdc00 || low >= 0xdfff) 833 Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); 834 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000; 835 } 836 if (uv < 0x10000) { 837 *d++ = (U8)(( uv >> 12) | 0xe0); 838 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); 839 *d++ = (U8)(( uv & 0x3f) | 0x80); 840 continue; 841 } 842 else { 843 *d++ = (U8)(( uv >> 18) | 0xf0); 844 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); 845 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); 846 *d++ = (U8)(( uv & 0x3f) | 0x80); 847 continue; 848 } 849 } 850 *newlen = d - dstart; 851 return d; 852 } 853 854 /* Note: this one is slightly destructive of the source. */ 855 856 U8* 857 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) 858 { 859 U8* s = (U8*)p; 860 U8* send = s + bytelen; 861 while (s < send) { 862 U8 tmp = s[0]; 863 s[0] = s[1]; 864 s[1] = tmp; 865 s += 2; 866 } 867 return utf16_to_utf8(p, d, bytelen, newlen); 868 } 869 870 /* for now these are all defined (inefficiently) in terms of the utf8 versions */ 871 872 bool 873 Perl_is_uni_alnum(pTHX_ UV c) 874 { 875 U8 tmpbuf[UTF8_MAXLEN+1]; 876 uvchr_to_utf8(tmpbuf, c); 877 return is_utf8_alnum(tmpbuf); 878 } 879 880 bool 881 Perl_is_uni_alnumc(pTHX_ UV c) 882 { 883 U8 tmpbuf[UTF8_MAXLEN+1]; 884 uvchr_to_utf8(tmpbuf, c); 885 return is_utf8_alnumc(tmpbuf); 886 } 887 888 bool 889 Perl_is_uni_idfirst(pTHX_ UV c) 890 { 891 U8 tmpbuf[UTF8_MAXLEN+1]; 892 uvchr_to_utf8(tmpbuf, c); 893 return is_utf8_idfirst(tmpbuf); 894 } 895 896 bool 897 Perl_is_uni_alpha(pTHX_ UV c) 898 { 899 U8 tmpbuf[UTF8_MAXLEN+1]; 900 uvchr_to_utf8(tmpbuf, c); 901 return is_utf8_alpha(tmpbuf); 902 } 903 904 bool 905 Perl_is_uni_ascii(pTHX_ UV c) 906 { 907 U8 tmpbuf[UTF8_MAXLEN+1]; 908 uvchr_to_utf8(tmpbuf, c); 909 return is_utf8_ascii(tmpbuf); 910 } 911 912 bool 913 Perl_is_uni_space(pTHX_ UV c) 914 { 915 U8 tmpbuf[UTF8_MAXLEN+1]; 916 uvchr_to_utf8(tmpbuf, c); 917 return is_utf8_space(tmpbuf); 918 } 919 920 bool 921 Perl_is_uni_digit(pTHX_ UV c) 922 { 923 U8 tmpbuf[UTF8_MAXLEN+1]; 924 uvchr_to_utf8(tmpbuf, c); 925 return is_utf8_digit(tmpbuf); 926 } 927 928 bool 929 Perl_is_uni_upper(pTHX_ UV c) 930 { 931 U8 tmpbuf[UTF8_MAXLEN+1]; 932 uvchr_to_utf8(tmpbuf, c); 933 return is_utf8_upper(tmpbuf); 934 } 935 936 bool 937 Perl_is_uni_lower(pTHX_ UV c) 938 { 939 U8 tmpbuf[UTF8_MAXLEN+1]; 940 uvchr_to_utf8(tmpbuf, c); 941 return is_utf8_lower(tmpbuf); 942 } 943 944 bool 945 Perl_is_uni_cntrl(pTHX_ UV c) 946 { 947 U8 tmpbuf[UTF8_MAXLEN+1]; 948 uvchr_to_utf8(tmpbuf, c); 949 return is_utf8_cntrl(tmpbuf); 950 } 951 952 bool 953 Perl_is_uni_graph(pTHX_ UV c) 954 { 955 U8 tmpbuf[UTF8_MAXLEN+1]; 956 uvchr_to_utf8(tmpbuf, c); 957 return is_utf8_graph(tmpbuf); 958 } 959 960 bool 961 Perl_is_uni_print(pTHX_ UV c) 962 { 963 U8 tmpbuf[UTF8_MAXLEN+1]; 964 uvchr_to_utf8(tmpbuf, c); 965 return is_utf8_print(tmpbuf); 966 } 967 968 bool 969 Perl_is_uni_punct(pTHX_ UV c) 970 { 971 U8 tmpbuf[UTF8_MAXLEN+1]; 972 uvchr_to_utf8(tmpbuf, c); 973 return is_utf8_punct(tmpbuf); 974 } 975 976 bool 977 Perl_is_uni_xdigit(pTHX_ UV c) 978 { 979 U8 tmpbuf[UTF8_MAXLEN_UCLC+1]; 980 uvchr_to_utf8(tmpbuf, c); 981 return is_utf8_xdigit(tmpbuf); 982 } 983 984 UV 985 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp) 986 { 987 uvchr_to_utf8(p, c); 988 return to_utf8_upper(p, p, lenp); 989 } 990 991 UV 992 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp) 993 { 994 uvchr_to_utf8(p, c); 995 return to_utf8_title(p, p, lenp); 996 } 997 998 UV 999 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp) 1000 { 1001 uvchr_to_utf8(p, c); 1002 return to_utf8_lower(p, p, lenp); 1003 } 1004 1005 UV 1006 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp) 1007 { 1008 uvchr_to_utf8(p, c); 1009 return to_utf8_fold(p, p, lenp); 1010 } 1011 1012 /* for now these all assume no locale info available for Unicode > 255 */ 1013 1014 bool 1015 Perl_is_uni_alnum_lc(pTHX_ UV c) 1016 { 1017 return is_uni_alnum(c); /* XXX no locale support yet */ 1018 } 1019 1020 bool 1021 Perl_is_uni_alnumc_lc(pTHX_ UV c) 1022 { 1023 return is_uni_alnumc(c); /* XXX no locale support yet */ 1024 } 1025 1026 bool 1027 Perl_is_uni_idfirst_lc(pTHX_ UV c) 1028 { 1029 return is_uni_idfirst(c); /* XXX no locale support yet */ 1030 } 1031 1032 bool 1033 Perl_is_uni_alpha_lc(pTHX_ UV c) 1034 { 1035 return is_uni_alpha(c); /* XXX no locale support yet */ 1036 } 1037 1038 bool 1039 Perl_is_uni_ascii_lc(pTHX_ UV c) 1040 { 1041 return is_uni_ascii(c); /* XXX no locale support yet */ 1042 } 1043 1044 bool 1045 Perl_is_uni_space_lc(pTHX_ UV c) 1046 { 1047 return is_uni_space(c); /* XXX no locale support yet */ 1048 } 1049 1050 bool 1051 Perl_is_uni_digit_lc(pTHX_ UV c) 1052 { 1053 return is_uni_digit(c); /* XXX no locale support yet */ 1054 } 1055 1056 bool 1057 Perl_is_uni_upper_lc(pTHX_ UV c) 1058 { 1059 return is_uni_upper(c); /* XXX no locale support yet */ 1060 } 1061 1062 bool 1063 Perl_is_uni_lower_lc(pTHX_ UV c) 1064 { 1065 return is_uni_lower(c); /* XXX no locale support yet */ 1066 } 1067 1068 bool 1069 Perl_is_uni_cntrl_lc(pTHX_ UV c) 1070 { 1071 return is_uni_cntrl(c); /* XXX no locale support yet */ 1072 } 1073 1074 bool 1075 Perl_is_uni_graph_lc(pTHX_ UV c) 1076 { 1077 return is_uni_graph(c); /* XXX no locale support yet */ 1078 } 1079 1080 bool 1081 Perl_is_uni_print_lc(pTHX_ UV c) 1082 { 1083 return is_uni_print(c); /* XXX no locale support yet */ 1084 } 1085 1086 bool 1087 Perl_is_uni_punct_lc(pTHX_ UV c) 1088 { 1089 return is_uni_punct(c); /* XXX no locale support yet */ 1090 } 1091 1092 bool 1093 Perl_is_uni_xdigit_lc(pTHX_ UV c) 1094 { 1095 return is_uni_xdigit(c); /* XXX no locale support yet */ 1096 } 1097 1098 U32 1099 Perl_to_uni_upper_lc(pTHX_ U32 c) 1100 { 1101 /* XXX returns only the first character -- do not use XXX */ 1102 /* XXX no locale support yet */ 1103 STRLEN len; 1104 U8 tmpbuf[UTF8_MAXLEN_UCLC+1]; 1105 return (U32)to_uni_upper(c, tmpbuf, &len); 1106 } 1107 1108 U32 1109 Perl_to_uni_title_lc(pTHX_ U32 c) 1110 { 1111 /* XXX returns only the first character XXX -- do not use XXX */ 1112 /* XXX no locale support yet */ 1113 STRLEN len; 1114 U8 tmpbuf[UTF8_MAXLEN_UCLC+1]; 1115 return (U32)to_uni_title(c, tmpbuf, &len); 1116 } 1117 1118 U32 1119 Perl_to_uni_lower_lc(pTHX_ U32 c) 1120 { 1121 /* XXX returns only the first character -- do not use XXX */ 1122 /* XXX no locale support yet */ 1123 STRLEN len; 1124 U8 tmpbuf[UTF8_MAXLEN_UCLC+1]; 1125 return (U32)to_uni_lower(c, tmpbuf, &len); 1126 } 1127 1128 bool 1129 Perl_is_utf8_alnum(pTHX_ U8 *p) 1130 { 1131 if (!is_utf8_char(p)) 1132 return FALSE; 1133 if (!PL_utf8_alnum) 1134 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true 1135 * descendant of isalnum(3), in other words, it doesn't 1136 * contain the '_'. --jhi */ 1137 PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0); 1138 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0; 1139 /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */ 1140 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */ 1141 if (!PL_utf8_alnum) 1142 PL_utf8_alnum = swash_init("utf8", "", 1143 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0); 1144 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0; 1145 #endif 1146 } 1147 1148 bool 1149 Perl_is_utf8_alnumc(pTHX_ U8 *p) 1150 { 1151 if (!is_utf8_char(p)) 1152 return FALSE; 1153 if (!PL_utf8_alnum) 1154 PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0); 1155 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0; 1156 /* return is_utf8_alpha(p) || is_utf8_digit(p); */ 1157 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */ 1158 if (!PL_utf8_alnum) 1159 PL_utf8_alnum = swash_init("utf8", "", 1160 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0); 1161 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0; 1162 #endif 1163 } 1164 1165 bool 1166 Perl_is_utf8_idfirst(pTHX_ U8 *p) /* The naming is historical. */ 1167 { 1168 if (*p == '_') 1169 return TRUE; 1170 if (!is_utf8_char(p)) 1171 return FALSE; 1172 if (!PL_utf8_idstart) /* is_utf8_idstart would be more logical. */ 1173 PL_utf8_idstart = swash_init("utf8", "IdStart", &PL_sv_undef, 0, 0); 1174 return swash_fetch(PL_utf8_idstart, p, TRUE) != 0; 1175 } 1176 1177 bool 1178 Perl_is_utf8_idcont(pTHX_ U8 *p) 1179 { 1180 if (*p == '_') 1181 return TRUE; 1182 if (!is_utf8_char(p)) 1183 return FALSE; 1184 if (!PL_utf8_idcont) 1185 PL_utf8_idcont = swash_init("utf8", "IdContinue", &PL_sv_undef, 0, 0); 1186 return swash_fetch(PL_utf8_idcont, p, TRUE) != 0; 1187 } 1188 1189 bool 1190 Perl_is_utf8_alpha(pTHX_ U8 *p) 1191 { 1192 if (!is_utf8_char(p)) 1193 return FALSE; 1194 if (!PL_utf8_alpha) 1195 PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0); 1196 return swash_fetch(PL_utf8_alpha, p, TRUE) != 0; 1197 } 1198 1199 bool 1200 Perl_is_utf8_ascii(pTHX_ U8 *p) 1201 { 1202 if (!is_utf8_char(p)) 1203 return FALSE; 1204 if (!PL_utf8_ascii) 1205 PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0); 1206 return swash_fetch(PL_utf8_ascii, p, TRUE) != 0; 1207 } 1208 1209 bool 1210 Perl_is_utf8_space(pTHX_ U8 *p) 1211 { 1212 if (!is_utf8_char(p)) 1213 return FALSE; 1214 if (!PL_utf8_space) 1215 PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0); 1216 return swash_fetch(PL_utf8_space, p, TRUE) != 0; 1217 } 1218 1219 bool 1220 Perl_is_utf8_digit(pTHX_ U8 *p) 1221 { 1222 if (!is_utf8_char(p)) 1223 return FALSE; 1224 if (!PL_utf8_digit) 1225 PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0); 1226 return swash_fetch(PL_utf8_digit, p, TRUE) != 0; 1227 } 1228 1229 bool 1230 Perl_is_utf8_upper(pTHX_ U8 *p) 1231 { 1232 if (!is_utf8_char(p)) 1233 return FALSE; 1234 if (!PL_utf8_upper) 1235 PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0); 1236 return swash_fetch(PL_utf8_upper, p, TRUE) != 0; 1237 } 1238 1239 bool 1240 Perl_is_utf8_lower(pTHX_ U8 *p) 1241 { 1242 if (!is_utf8_char(p)) 1243 return FALSE; 1244 if (!PL_utf8_lower) 1245 PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0); 1246 return swash_fetch(PL_utf8_lower, p, TRUE) != 0; 1247 } 1248 1249 bool 1250 Perl_is_utf8_cntrl(pTHX_ U8 *p) 1251 { 1252 if (!is_utf8_char(p)) 1253 return FALSE; 1254 if (!PL_utf8_cntrl) 1255 PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0); 1256 return swash_fetch(PL_utf8_cntrl, p, TRUE) != 0; 1257 } 1258 1259 bool 1260 Perl_is_utf8_graph(pTHX_ U8 *p) 1261 { 1262 if (!is_utf8_char(p)) 1263 return FALSE; 1264 if (!PL_utf8_graph) 1265 PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0); 1266 return swash_fetch(PL_utf8_graph, p, TRUE) != 0; 1267 } 1268 1269 bool 1270 Perl_is_utf8_print(pTHX_ U8 *p) 1271 { 1272 if (!is_utf8_char(p)) 1273 return FALSE; 1274 if (!PL_utf8_print) 1275 PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0); 1276 return swash_fetch(PL_utf8_print, p, TRUE) != 0; 1277 } 1278 1279 bool 1280 Perl_is_utf8_punct(pTHX_ U8 *p) 1281 { 1282 if (!is_utf8_char(p)) 1283 return FALSE; 1284 if (!PL_utf8_punct) 1285 PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0); 1286 return swash_fetch(PL_utf8_punct, p, TRUE) != 0; 1287 } 1288 1289 bool 1290 Perl_is_utf8_xdigit(pTHX_ U8 *p) 1291 { 1292 if (!is_utf8_char(p)) 1293 return FALSE; 1294 if (!PL_utf8_xdigit) 1295 PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0); 1296 return swash_fetch(PL_utf8_xdigit, p, TRUE) != 0; 1297 } 1298 1299 bool 1300 Perl_is_utf8_mark(pTHX_ U8 *p) 1301 { 1302 if (!is_utf8_char(p)) 1303 return FALSE; 1304 if (!PL_utf8_mark) 1305 PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0); 1306 return swash_fetch(PL_utf8_mark, p, TRUE) != 0; 1307 } 1308 1309 /* 1310 =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special 1311 1312 The "p" contains the pointer to the UTF-8 string encoding 1313 the character that is being converted. 1314 1315 The "ustrp" is a pointer to the character buffer to put the 1316 conversion result to. The "lenp" is a pointer to the length 1317 of the result. 1318 1319 The "swashp" is a pointer to the swash to use. 1320 1321 Both the special and normal mappings are stored lib/unicore/To/Foo.pl, 1322 and loaded by SWASHGET, using lib/utf8_heavy.pl. The special (usually, 1323 but not always, a multicharacter mapping), is tried first. 1324 1325 The "special" is a string like "utf8::ToSpecLower", which means the 1326 hash %utf8::ToSpecLower. The access to the hash is through 1327 Perl_to_utf8_case(). 1328 1329 The "normal" is a string like "ToLower" which means the swash 1330 %utf8::ToLower. 1331 1332 =cut */ 1333 1334 UV 1335 Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp, char *normal, char *special) 1336 { 1337 UV uv0, uv1; 1338 U8 tmpbuf[UTF8_MAXLEN_FOLD+1]; 1339 STRLEN len = 0; 1340 1341 uv0 = utf8_to_uvchr(p, 0); 1342 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings 1343 * are necessary in EBCDIC, they are redundant no-ops 1344 * in ASCII-ish platforms, and hopefully optimized away. */ 1345 uv1 = NATIVE_TO_UNI(uv0); 1346 uvuni_to_utf8(tmpbuf, uv1); 1347 1348 if (!*swashp) /* load on-demand */ 1349 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0); 1350 1351 if (special) { 1352 /* It might be "special" (sometimes, but not always, 1353 * a multicharacter mapping) */ 1354 HV *hv; 1355 SV *keysv; 1356 HE *he; 1357 SV *val; 1358 1359 if ((hv = get_hv(special, FALSE)) && 1360 (keysv = sv_2mortal(Perl_newSVpvf(aTHX_ "%04"UVXf, uv1))) && 1361 (he = hv_fetch_ent(hv, keysv, FALSE, 0)) && 1362 (val = HeVAL(he))) { 1363 char *s; 1364 1365 s = SvPV(val, len); 1366 if (len == 1) 1367 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp; 1368 else { 1369 #ifdef EBCDIC 1370 /* If we have EBCDIC we need to remap the characters 1371 * since any characters in the low 256 are Unicode 1372 * code points, not EBCDIC. */ 1373 U8 *t = (U8*)s, *tend = t + len, *d; 1374 1375 d = tmpbuf; 1376 if (SvUTF8(val)) { 1377 STRLEN tlen = 0; 1378 1379 while (t < tend) { 1380 UV c = utf8_to_uvchr(t, &tlen); 1381 if (tlen > 0) { 1382 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c)); 1383 t += tlen; 1384 } 1385 else 1386 break; 1387 } 1388 } 1389 else { 1390 while (t < tend) { 1391 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t)); 1392 t++; 1393 } 1394 } 1395 len = d - tmpbuf; 1396 Copy(tmpbuf, ustrp, len, U8); 1397 #else 1398 Copy(s, ustrp, len, U8); 1399 #endif 1400 } 1401 } 1402 } 1403 1404 if (!len && *swashp) { 1405 UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE); 1406 1407 if (uv2) { 1408 /* It was "normal" (a single character mapping). */ 1409 UV uv3 = UNI_TO_NATIVE(uv2); 1410 1411 len = uvchr_to_utf8(ustrp, uv3) - ustrp; 1412 } 1413 } 1414 1415 if (!len) /* Neither: just copy. */ 1416 len = uvchr_to_utf8(ustrp, uv0) - ustrp; 1417 1418 if (lenp) 1419 *lenp = len; 1420 1421 return len ? utf8_to_uvchr(ustrp, 0) : 0; 1422 } 1423 1424 /* 1425 =for apidoc A|UV|to_utf8_upper|U8 *p|U8 *ustrp|STRLEN *lenp 1426 1427 Convert the UTF-8 encoded character at p to its uppercase version and 1428 store that in UTF-8 in ustrp and its length in bytes in lenp. Note 1429 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the 1430 uppercase version may be longer than the original character (up to two 1431 characters). 1432 1433 The first character of the uppercased version is returned 1434 (but note, as explained above, that there may be more.) 1435 1436 =cut */ 1437 1438 UV 1439 Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) 1440 { 1441 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, 1442 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); 1443 } 1444 1445 /* 1446 =for apidoc A|UV|to_utf8_title|U8 *p|U8 *ustrp|STRLEN *lenp 1447 1448 Convert the UTF-8 encoded character at p to its titlecase version and 1449 store that in UTF-8 in ustrp and its length in bytes in lenp. Note 1450 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the 1451 titlecase version may be longer than the original character (up to two 1452 characters). 1453 1454 The first character of the titlecased version is returned 1455 (but note, as explained above, that there may be more.) 1456 1457 =cut */ 1458 1459 UV 1460 Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) 1461 { 1462 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, 1463 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle"); 1464 } 1465 1466 /* 1467 =for apidoc A|UV|to_utf8_lower|U8 *p|U8 *ustrp|STRLEN *lenp 1468 1469 Convert the UTF-8 encoded character at p to its lowercase version and 1470 store that in UTF-8 in ustrp and its length in bytes in lenp. Note 1471 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the 1472 lowercase version may be longer than the original character (up to two 1473 characters). 1474 1475 The first character of the lowercased version is returned 1476 (but note, as explained above, that there may be more.) 1477 1478 =cut */ 1479 1480 UV 1481 Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) 1482 { 1483 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, 1484 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); 1485 } 1486 1487 /* 1488 =for apidoc A|UV|to_utf8_fold|U8 *p|U8 *ustrp|STRLEN *lenp 1489 1490 Convert the UTF-8 encoded character at p to its foldcase version and 1491 store that in UTF-8 in ustrp and its length in bytes in lenp. Note 1492 that the ustrp needs to be at least UTF8_MAXLEN_FOLD+1 bytes since the 1493 foldcase version may be longer than the original character (up to 1494 three characters). 1495 1496 The first character of the foldcased version is returned 1497 (but note, as explained above, that there may be more.) 1498 1499 =cut */ 1500 1501 UV 1502 Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) 1503 { 1504 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, 1505 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold"); 1506 } 1507 1508 /* a "swash" is a swatch hash */ 1509 1510 SV* 1511 Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none) 1512 { 1513 SV* retval; 1514 SV* tokenbufsv = sv_2mortal(NEWSV(0,0)); 1515 dSP; 1516 HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE); 1517 SV* errsv_save; 1518 1519 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */ 1520 ENTER; 1521 errsv_save = newSVsv(ERRSV); 1522 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv); 1523 if (!SvTRUE(ERRSV)) 1524 sv_setsv(ERRSV, errsv_save); 1525 SvREFCNT_dec(errsv_save); 1526 LEAVE; 1527 } 1528 SPAGAIN; 1529 PUSHSTACKi(PERLSI_MAGIC); 1530 PUSHMARK(SP); 1531 EXTEND(SP,5); 1532 PUSHs(sv_2mortal(newSVpvn(pkg, strlen(pkg)))); 1533 PUSHs(sv_2mortal(newSVpvn(name, strlen(name)))); 1534 PUSHs(listsv); 1535 PUSHs(sv_2mortal(newSViv(minbits))); 1536 PUSHs(sv_2mortal(newSViv(none))); 1537 PUTBACK; 1538 ENTER; 1539 SAVEI32(PL_hints); 1540 PL_hints = 0; 1541 save_re_context(); 1542 if (PL_curcop == &PL_compiling) { 1543 /* XXX ought to be handled by lex_start */ 1544 SAVEI32(PL_in_my); 1545 sv_setpv(tokenbufsv, PL_tokenbuf); 1546 } 1547 errsv_save = newSVsv(ERRSV); 1548 if (call_method("SWASHNEW", G_SCALAR)) 1549 retval = newSVsv(*PL_stack_sp--); 1550 else 1551 retval = &PL_sv_undef; 1552 if (!SvTRUE(ERRSV)) 1553 sv_setsv(ERRSV, errsv_save); 1554 SvREFCNT_dec(errsv_save); 1555 LEAVE; 1556 POPSTACK; 1557 if (PL_curcop == &PL_compiling) { 1558 STRLEN len; 1559 char* pv = SvPV(tokenbufsv, len); 1560 1561 Copy(pv, PL_tokenbuf, len+1, char); 1562 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK); 1563 } 1564 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) { 1565 if (SvPOK(retval)) 1566 Perl_croak(aTHX_ "Can't find Unicode property definition \"%s\"", 1567 SvPV_nolen(retval)); 1568 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); 1569 } 1570 return retval; 1571 } 1572 1573 1574 /* This API is wrong for special case conversions since we may need to 1575 * return several Unicode characters for a single Unicode character 1576 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is 1577 * the lower-level routine, and it is similarly broken for returning 1578 * multiple values. --jhi */ 1579 UV 1580 Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8) 1581 { 1582 HV* hv = (HV*)SvRV(sv); 1583 U32 klen; 1584 U32 off; 1585 STRLEN slen; 1586 STRLEN needents; 1587 U8 *tmps = NULL; 1588 U32 bit; 1589 SV *retval; 1590 U8 tmputf8[2]; 1591 UV c = NATIVE_TO_ASCII(*ptr); 1592 1593 if (!do_utf8 && !UNI_IS_INVARIANT(c)) { 1594 tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c); 1595 tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c); 1596 ptr = tmputf8; 1597 } 1598 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ 1599 * then the "swatch" is a vec() for al the chars which start 1600 * with 0xAA..0xYY 1601 * So the key in the hash (klen) is length of encoded char -1 1602 */ 1603 klen = UTF8SKIP(ptr) - 1; 1604 off = ptr[klen]; 1605 1606 if (klen == 0) 1607 { 1608 /* If char in invariant then swatch is for all the invariant chars 1609 * In both UTF-8 and UTF8-MOD that happens to be UTF_CONTINUATION_MARK 1610 */ 1611 needents = UTF_CONTINUATION_MARK; 1612 off = NATIVE_TO_UTF(ptr[klen]); 1613 } 1614 else 1615 { 1616 /* If char is encoded then swatch is for the prefix */ 1617 needents = (1 << UTF_ACCUMULATION_SHIFT); 1618 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK; 1619 } 1620 1621 /* 1622 * This single-entry cache saves about 1/3 of the utf8 overhead in test 1623 * suite. (That is, only 7-8% overall over just a hash cache. Still, 1624 * it's nothing to sniff at.) Pity we usually come through at least 1625 * two function calls to get here... 1626 * 1627 * NB: this code assumes that swatches are never modified, once generated! 1628 */ 1629 1630 if (hv == PL_last_swash_hv && 1631 klen == PL_last_swash_klen && 1632 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) 1633 { 1634 tmps = PL_last_swash_tmps; 1635 slen = PL_last_swash_slen; 1636 } 1637 else { 1638 /* Try our second-level swatch cache, kept in a hash. */ 1639 SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE); 1640 1641 /* If not cached, generate it via utf8::SWASHGET */ 1642 if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) { 1643 dSP; 1644 /* We use utf8n_to_uvuni() as we want an index into 1645 Unicode tables, not a native character number. 1646 */ 1647 UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, 0, 1648 ckWARN(WARN_UTF8) ? 1649 0 : UTF8_ALLOW_ANY); 1650 SV *errsv_save; 1651 ENTER; 1652 SAVETMPS; 1653 save_re_context(); 1654 PUSHSTACKi(PERLSI_MAGIC); 1655 PUSHMARK(SP); 1656 EXTEND(SP,3); 1657 PUSHs((SV*)sv); 1658 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */ 1659 PUSHs(sv_2mortal(newSViv((klen) ? 1660 (code_point & ~(needents - 1)) : 0))); 1661 PUSHs(sv_2mortal(newSViv(needents))); 1662 PUTBACK; 1663 errsv_save = newSVsv(ERRSV); 1664 if (call_method("SWASHGET", G_SCALAR)) 1665 retval = newSVsv(*PL_stack_sp--); 1666 else 1667 retval = &PL_sv_undef; 1668 if (!SvTRUE(ERRSV)) 1669 sv_setsv(ERRSV, errsv_save); 1670 SvREFCNT_dec(errsv_save); 1671 POPSTACK; 1672 FREETMPS; 1673 LEAVE; 1674 if (PL_curcop == &PL_compiling) 1675 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK); 1676 1677 svp = hv_store(hv, (char*)ptr, klen, retval, 0); 1678 1679 if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents) 1680 Perl_croak(aTHX_ "SWASHGET didn't return result of proper length"); 1681 } 1682 1683 PL_last_swash_hv = hv; 1684 PL_last_swash_klen = klen; 1685 PL_last_swash_tmps = tmps; 1686 PL_last_swash_slen = slen; 1687 if (klen) 1688 Copy(ptr, PL_last_swash_key, klen, U8); 1689 } 1690 1691 switch ((int)((slen << 3) / needents)) { 1692 case 1: 1693 bit = 1 << (off & 7); 1694 off >>= 3; 1695 return (tmps[off] & bit) != 0; 1696 case 8: 1697 return tmps[off]; 1698 case 16: 1699 off <<= 1; 1700 return (tmps[off] << 8) + tmps[off + 1] ; 1701 case 32: 1702 off <<= 2; 1703 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; 1704 } 1705 Perl_croak(aTHX_ "panic: swash_fetch"); 1706 return 0; 1707 } 1708 1709 1710 /* 1711 =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv 1712 1713 Adds the UTF8 representation of the Native codepoint C<uv> to the end 1714 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free 1715 bytes available. The return value is the pointer to the byte after the 1716 end of the new character. In other words, 1717 1718 d = uvchr_to_utf8(d, uv); 1719 1720 is the recommended wide native character-aware way of saying 1721 1722 *(d++) = uv; 1723 1724 =cut 1725 */ 1726 1727 /* On ASCII machines this is normally a macro but we want a 1728 real function in case XS code wants it 1729 */ 1730 #undef Perl_uvchr_to_utf8 1731 U8 * 1732 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) 1733 { 1734 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0); 1735 } 1736 1737 U8 * 1738 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) 1739 { 1740 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags); 1741 } 1742 1743 /* 1744 =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags 1745 1746 Returns the native character value of the first character in the string C<s> 1747 which is assumed to be in UTF8 encoding; C<retlen> will be set to the 1748 length, in bytes, of that character. 1749 1750 Allows length and flags to be passed to low level routine. 1751 1752 =cut 1753 */ 1754 /* On ASCII machines this is normally a macro but we want 1755 a real function in case XS code wants it 1756 */ 1757 #undef Perl_utf8n_to_uvchr 1758 UV 1759 Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) 1760 { 1761 UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags); 1762 return UNI_TO_NATIVE(uv); 1763 } 1764 1765 /* 1766 =for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags 1767 1768 Build to the scalar dsv a displayable version of the string spv, 1769 length len, the displayable version being at most pvlim bytes long 1770 (if longer, the rest is truncated and "..." will be appended). 1771 1772 The flags argument can have UNI_DISPLAY_ISPRINT set to display 1773 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH 1774 to display the \\[nrfta\\] as the backslashed versions (like '\n') 1775 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\). 1776 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both 1777 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on. 1778 1779 The pointer to the PV of the dsv is returned. 1780 1781 =cut */ 1782 char * 1783 Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags) 1784 { 1785 int truncated = 0; 1786 char *s, *e; 1787 1788 sv_setpvn(dsv, "", 0); 1789 for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) { 1790 UV u; 1791 bool ok = FALSE; 1792 1793 if (pvlim && SvCUR(dsv) >= pvlim) { 1794 truncated++; 1795 break; 1796 } 1797 u = utf8_to_uvchr((U8*)s, 0); 1798 if (u < 256) { 1799 if (!ok && (flags & UNI_DISPLAY_BACKSLASH)) { 1800 switch (u & 0xFF) { 1801 case '\n': 1802 Perl_sv_catpvf(aTHX_ dsv, "\\n"); ok = TRUE; break; 1803 case '\r': 1804 Perl_sv_catpvf(aTHX_ dsv, "\\r"); ok = TRUE; break; 1805 case '\t': 1806 Perl_sv_catpvf(aTHX_ dsv, "\\t"); ok = TRUE; break; 1807 case '\f': 1808 Perl_sv_catpvf(aTHX_ dsv, "\\f"); ok = TRUE; break; 1809 case '\a': 1810 Perl_sv_catpvf(aTHX_ dsv, "\\a"); ok = TRUE; break; 1811 case '\\': 1812 Perl_sv_catpvf(aTHX_ dsv, "\\\\" ); ok = TRUE; break; 1813 default: break; 1814 } 1815 } 1816 /* isPRINT() is the locale-blind version. */ 1817 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(u & 0xFF)) { 1818 Perl_sv_catpvf(aTHX_ dsv, "%c", (char)(u & 0xFF)); 1819 ok = TRUE; 1820 } 1821 } 1822 if (!ok) 1823 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u); 1824 } 1825 if (truncated) 1826 sv_catpvn(dsv, "...", 3); 1827 1828 return SvPVX(dsv); 1829 } 1830 1831 /* 1832 =for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags 1833 1834 Build to the scalar dsv a displayable version of the scalar sv, 1835 the displayable version being at most pvlim bytes long 1836 (if longer, the rest is truncated and "..." will be appended). 1837 1838 The flags argument is as in pv_uni_display(). 1839 1840 The pointer to the PV of the dsv is returned. 1841 1842 =cut */ 1843 char * 1844 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags) 1845 { 1846 return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv), 1847 pvlim, flags); 1848 } 1849 1850 /* 1851 =for apidoc A|I32|ibcmp_utf8|const char *s1|char **pe1|register UV l1|bool u1|const char *s2|char **pe2|register UV l2|bool u2 1852 1853 Return true if the strings s1 and s2 differ case-insensitively, false 1854 if not (if they are equal case-insensitively). If u1 is true, the 1855 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true, 1856 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2 1857 are false, the respective string is assumed to be in native 8-bit 1858 encoding. 1859 1860 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied 1861 in there (they will point at the beginning of the I<next> character). 1862 If the pointers behind pe1 or pe2 are non-NULL, they are the end 1863 pointers beyond which scanning will not continue under any 1864 circustances. If the byte lengths l1 and l2 are non-zero, s1+l1 and 1865 s2+l2 will be used as goal end pointers that will also stop the scan, 1866 and which qualify towards defining a successful match: all the scans 1867 that define an explicit length must reach their goal pointers for 1868 a match to succeed). 1869 1870 For case-insensitiveness, the "casefolding" of Unicode is used 1871 instead of upper/lowercasing both the characters, see 1872 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings). 1873 1874 =cut */ 1875 I32 1876 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2) 1877 { 1878 register U8 *p1 = (U8*)s1; 1879 register U8 *p2 = (U8*)s2; 1880 register U8 *e1 = 0, *f1 = 0, *q1 = 0; 1881 register U8 *e2 = 0, *f2 = 0, *q2 = 0; 1882 STRLEN n1 = 0, n2 = 0; 1883 U8 foldbuf1[UTF8_MAXLEN_FOLD+1]; 1884 U8 foldbuf2[UTF8_MAXLEN_FOLD+1]; 1885 U8 natbuf[1+1]; 1886 STRLEN foldlen1, foldlen2; 1887 bool match; 1888 1889 if (pe1) 1890 e1 = *(U8**)pe1; 1891 if (e1 == 0 || (l1 && l1 < (UV)(e1 - (U8*)s1))) 1892 f1 = (U8*)s1 + l1; 1893 if (pe2) 1894 e2 = *(U8**)pe2; 1895 if (e2 == 0 || (l2 && l2 < (UV)(e2 - (U8*)s2))) 1896 f2 = (U8*)s2 + l2; 1897 1898 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)) 1899 return 1; /* mismatch; possible infinite loop or false positive */ 1900 1901 if (!u1 || !u2) 1902 natbuf[1] = 0; /* Need to terminate the buffer. */ 1903 1904 while ((e1 == 0 || p1 < e1) && 1905 (f1 == 0 || p1 < f1) && 1906 (e2 == 0 || p2 < e2) && 1907 (f2 == 0 || p2 < f2)) { 1908 if (n1 == 0) { 1909 if (u1) 1910 to_utf8_fold(p1, foldbuf1, &foldlen1); 1911 else { 1912 natbuf[0] = *p1; 1913 to_utf8_fold(natbuf, foldbuf1, &foldlen1); 1914 } 1915 q1 = foldbuf1; 1916 n1 = foldlen1; 1917 } 1918 if (n2 == 0) { 1919 if (u2) 1920 to_utf8_fold(p2, foldbuf2, &foldlen2); 1921 else { 1922 natbuf[0] = *p2; 1923 to_utf8_fold(natbuf, foldbuf2, &foldlen2); 1924 } 1925 q2 = foldbuf2; 1926 n2 = foldlen2; 1927 } 1928 while (n1 && n2) { 1929 if ( UTF8SKIP(q1) != UTF8SKIP(q2) || 1930 (UTF8SKIP(q1) == 1 && *q1 != *q2) || 1931 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) ) 1932 return 1; /* mismatch */ 1933 n1 -= UTF8SKIP(q1); 1934 q1 += UTF8SKIP(q1); 1935 n2 -= UTF8SKIP(q2); 1936 q2 += UTF8SKIP(q2); 1937 } 1938 if (n1 == 0) 1939 p1 += u1 ? UTF8SKIP(p1) : 1; 1940 if (n2 == 0) 1941 p2 += u2 ? UTF8SKIP(p2) : 1; 1942 1943 } 1944 1945 /* A match is defined by all the scans that specified 1946 * an explicit length reaching their final goals. */ 1947 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2); 1948 1949 if (match) { 1950 if (pe1) 1951 *pe1 = (char*)p1; 1952 if (pe2) 1953 *pe2 = (char*)p2; 1954 } 1955 1956 return match ? 0 : 1; /* 0 match, 1 mismatch */ 1957 } 1958 1959