bwstring.c revision 235987
1/*- 2 * Copyright (C) 2009 Gabor Kovesdan <gabor@FreeBSD.org> 3 * Copyright (C) 2012 Oleg Moskalenko <oleg.moskalenko@citrix.com> 4 * All rights reserved. 5 * 6 * Redistribution and use in source and binary forms, with or without 7 * modification, are permitted provided that the following conditions 8 * are met: 9 * 1. Redistributions of source code must retain the above copyright 10 * notice, this list of conditions and the following disclaimer. 11 * 2. Redistributions in binary form must reproduce the above copyright 12 * notice, this list of conditions and the following disclaimer in the 13 * documentation and/or other materials provided with the distribution. 14 * 15 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 16 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 17 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 18 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 19 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 20 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 21 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 22 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 23 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 24 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 25 * SUCH DAMAGE. 26 */ 27 28#include <sys/cdefs.h> 29__FBSDID("$FreeBSD: head/usr.bin/sort/bwstring.c 235987 2012-05-25 09:30:16Z gabor $"); 30 31#include <ctype.h> 32#include <errno.h> 33#include <err.h> 34#include <langinfo.h> 35#include <math.h> 36#include <stdlib.h> 37#include <string.h> 38#include <wchar.h> 39#include <wctype.h> 40 41#include "bwstring.h" 42#include "sort.h" 43 44bool byte_sort; 45 46static wchar_t **wmonths; 47static unsigned char **cmonths; 48 49/* initialise months */ 50 51void 52initialise_months(void) 53{ 54 const nl_item item[12] = { ABMON_1, ABMON_2, ABMON_3, ABMON_4, 55 ABMON_5, ABMON_6, ABMON_7, ABMON_8, ABMON_9, ABMON_10, 56 ABMON_11, ABMON_12 }; 57 unsigned char *tmp; 58 size_t len; 59 60 if (MB_CUR_MAX == 1) { 61 if (cmonths == NULL) { 62 unsigned char *m; 63 64 cmonths = sort_malloc(sizeof(unsigned char*) * 12); 65 for (int i = 0; i < 12; i++) { 66 cmonths[i] = NULL; 67 tmp = (unsigned char *) nl_langinfo(item[i]); 68 if (tmp == NULL) 69 continue; 70 if (debug_sort) 71 printf("month[%d]=%s\n", i, tmp); 72 len = strlen(tmp); 73 if (len < 1) 74 continue; 75 while (isblank(*tmp)) 76 ++tmp; 77 m = sort_malloc(len + 1); 78 memcpy(m, tmp, len + 1); 79 m[len] = '\0'; 80 for (unsigned int j = 0; j < len; j++) 81 m[j] = toupper(m[j]); 82 cmonths[i] = m; 83 } 84 } 85 86 } else { 87 if (wmonths == NULL) { 88 wchar_t *m; 89 90 wmonths = sort_malloc(sizeof(wchar_t *) * 12); 91 for (int i = 0; i < 12; i++) { 92 wmonths[i] = NULL; 93 tmp = (unsigned char *) nl_langinfo(item[i]); 94 if (tmp == NULL) 95 continue; 96 if (debug_sort) 97 printf("month[%d]=%s\n", i, tmp); 98 len = strlen(tmp); 99 if (len < 1) 100 continue; 101 while (isblank(*tmp)) 102 ++tmp; 103 m = sort_malloc(SIZEOF_WCHAR_STRING(len + 1)); 104 if (mbstowcs(m, tmp, len) == ((size_t) -1)) 105 continue; 106 m[len] = L'\0'; 107 for (unsigned int j = 0; j < len; j++) 108 m[j] = towupper(m[j]); 109 wmonths[i] = m; 110 } 111 } 112 } 113} 114 115/* 116 * Compare two wide-character strings 117 */ 118static int 119wide_str_coll(const wchar_t *s1, const wchar_t *s2) 120{ 121 int ret = 0; 122 123 errno = 0; 124 ret = wcscoll(s1, s2); 125 if (errno == EILSEQ) { 126 errno = 0; 127 ret = wcscmp(s1, s2); 128 if (errno != 0) { 129 for (size_t i = 0; ; ++i) { 130 wchar_t c1 = s1[i]; 131 wchar_t c2 = s2[i]; 132 if (c1 == L'\0') 133 return ((c2 == L'\0') ? 0 : -1); 134 if (c2 == L'\0') 135 return (+1); 136 if (c1 == c2) 137 continue; 138 return ((int)(c1 - c2)); 139 } 140 } 141 } 142 return (ret); 143} 144 145/* counterparts of wcs functions */ 146 147void 148bwsprintf(FILE *f, struct bwstring *bws, const char *prefix, const char *suffix) 149{ 150 151 if (MB_CUR_MAX == 1) 152 fprintf(f, "%s%s%s", prefix, bws->data.cstr, suffix); 153 else 154 fprintf(f, "%s%S%s", prefix, bws->data.wstr, suffix); 155} 156 157const void* bwsrawdata(const struct bwstring *bws) 158{ 159 160 return (&(bws->data)); 161} 162 163size_t bwsrawlen(const struct bwstring *bws) 164{ 165 166 return ((MB_CUR_MAX == 1) ? bws->len : SIZEOF_WCHAR_STRING(bws->len)); 167} 168 169size_t 170bws_memsize(const struct bwstring *bws) 171{ 172 173 return ((MB_CUR_MAX == 1) ? (bws->len + 2 + sizeof(struct bwstring)) : 174 (SIZEOF_WCHAR_STRING(bws->len + 1) + sizeof(struct bwstring))); 175} 176 177void 178bws_setlen(struct bwstring *bws, size_t newlen) 179{ 180 181 if (bws && newlen != bws->len && newlen <= bws->len) { 182 bws->len = newlen; 183 if (MB_CUR_MAX == 1) 184 bws->data.cstr[newlen] = '\0'; 185 else 186 bws->data.wstr[newlen] = L'\0'; 187 } 188} 189 190/* 191 * Allocate a new binary string of specified size 192 */ 193struct bwstring * 194bwsalloc(size_t sz) 195{ 196 struct bwstring *ret; 197 198 if (MB_CUR_MAX == 1) 199 ret = sort_malloc(sizeof(struct bwstring) + 1 + sz); 200 else 201 ret = sort_malloc(sizeof(struct bwstring) + 202 SIZEOF_WCHAR_STRING(sz + 1)); 203 ret->len = sz; 204 205 if (MB_CUR_MAX == 1) 206 ret->data.cstr[ret->len] = '\0'; 207 else 208 ret->data.wstr[ret->len] = L'\0'; 209 210 return (ret); 211} 212 213/* 214 * Create a copy of binary string. 215 * New string size equals the length of the old string. 216 */ 217struct bwstring * 218bwsdup(const struct bwstring *s) 219{ 220 221 if (s == NULL) 222 return (NULL); 223 else { 224 struct bwstring *ret = bwsalloc(s->len); 225 226 if (MB_CUR_MAX == 1) 227 memcpy(ret->data.cstr, s->data.cstr, (s->len)); 228 else 229 memcpy(ret->data.wstr, s->data.wstr, 230 SIZEOF_WCHAR_STRING(s->len)); 231 232 return (ret); 233 } 234} 235 236/* 237 * Create a new binary string from a raw binary buffer. 238 */ 239struct bwstring * 240bwssbdup(const wchar_t *str, size_t len) 241{ 242 243 if (str == NULL) 244 return ((len == 0) ? bwsalloc(0) : NULL); 245 else { 246 struct bwstring *ret; 247 248 ret = bwsalloc(len); 249 250 if (MB_CUR_MAX == 1) 251 for (size_t i = 0; i < len; ++i) 252 ret->data.cstr[i] = (unsigned char) str[i]; 253 else 254 memcpy(ret->data.wstr, str, SIZEOF_WCHAR_STRING(len)); 255 256 return (ret); 257 } 258} 259 260/* 261 * Create a new binary string from a raw binary buffer. 262 */ 263struct bwstring * 264bwscsbdup(const unsigned char *str, size_t len) 265{ 266 struct bwstring *ret; 267 268 ret = bwsalloc(len); 269 270 if (str) { 271 if (MB_CUR_MAX == 1) 272 memcpy(ret->data.cstr, str, len); 273 else { 274 mbstate_t mbs; 275 const char *s; 276 size_t charlen, chars, cptr; 277 278 charlen = chars = 0; 279 cptr = 0; 280 s = (const char *) str; 281 282 memset(&mbs, 0, sizeof(mbs)); 283 284 while (cptr < len) { 285 size_t n = MB_CUR_MAX; 286 287 if (n > len - cptr) 288 n = len - cptr; 289 charlen = mbrlen(s + cptr, n, &mbs); 290 switch (charlen) { 291 case 0: 292 /* FALLTHROUGH */ 293 case (size_t) -1: 294 /* FALLTHROUGH */ 295 case (size_t) -2: 296 ret->data.wstr[chars++] = 297 (unsigned char) s[cptr]; 298 ++cptr; 299 break; 300 default: 301 n = mbrtowc(ret->data.wstr + (chars++), 302 s + cptr, charlen, &mbs); 303 if ((n == (size_t)-1) || (n == (size_t)-2)) 304 /* NOTREACHED */ 305 err(2, "mbrtowc error"); 306 cptr += charlen; 307 }; 308 } 309 310 ret->len = chars; 311 ret->data.wstr[ret->len] = L'\0'; 312 } 313 } 314 return (ret); 315} 316 317/* 318 * De-allocate object memory 319 */ 320void 321bwsfree(const struct bwstring *s) 322{ 323 324 if (s) 325 sort_free(s); 326} 327 328/* 329 * Copy content of src binary string to dst. 330 * If the capacity of the dst string is not sufficient, 331 * then the data is truncated. 332 */ 333size_t 334bwscpy(struct bwstring *dst, const struct bwstring *src) 335{ 336 size_t nums = src->len; 337 338 if (nums > dst->len) 339 nums = dst->len; 340 dst->len = nums; 341 342 if (MB_CUR_MAX == 1) { 343 memcpy(dst->data.cstr, src->data.cstr, nums); 344 dst->data.cstr[dst->len] = '\0'; 345 } else { 346 memcpy(dst->data.wstr, src->data.wstr, 347 SIZEOF_WCHAR_STRING(nums + 1)); 348 dst->data.wstr[dst->len] = L'\0'; 349 } 350 351 return (nums); 352} 353 354/* 355 * Copy content of src binary string to dst, 356 * with specified number of symbols to be copied. 357 * If the capacity of the dst string is not sufficient, 358 * then the data is truncated. 359 */ 360struct bwstring * 361bwsncpy(struct bwstring *dst, const struct bwstring *src, size_t size) 362{ 363 size_t nums = src->len; 364 365 if (nums > dst->len) 366 nums = dst->len; 367 if (nums > size) 368 nums = size; 369 dst->len = nums; 370 371 if (MB_CUR_MAX == 1) { 372 memcpy(dst->data.cstr, src->data.cstr, nums); 373 dst->data.cstr[dst->len] = '\0'; 374 } else { 375 memcpy(dst->data.wstr, src->data.wstr, 376 SIZEOF_WCHAR_STRING(nums + 1)); 377 dst->data.wstr[dst->len] = L'\0'; 378 } 379 380 return (dst); 381} 382 383/* 384 * Copy content of src binary string to dst, 385 * with specified number of symbols to be copied. 386 * An offset value can be specified, from the start of src string. 387 * If the capacity of the dst string is not sufficient, 388 * then the data is truncated. 389 */ 390struct bwstring * 391bwsnocpy(struct bwstring *dst, const struct bwstring *src, size_t offset, 392 size_t size) 393{ 394 395 if (offset >= src->len) { 396 dst->data.wstr[0] = 0; 397 dst->len = 0; 398 } else { 399 size_t nums = src->len - offset; 400 401 if (nums > dst->len) 402 nums = dst->len; 403 if (nums > size) 404 nums = size; 405 dst->len = nums; 406 if (MB_CUR_MAX == 1) { 407 memcpy(dst->data.cstr, src->data.cstr + offset, 408 (nums)); 409 dst->data.cstr[dst->len] = '\0'; 410 } else { 411 memcpy(dst->data.wstr, src->data.wstr + offset, 412 SIZEOF_WCHAR_STRING(nums)); 413 dst->data.wstr[dst->len] = L'\0'; 414 } 415 } 416 return (dst); 417} 418 419/* 420 * Write binary string to the file. 421 * The output is ended either with '\n' (nl == true) 422 * or '\0' (nl == false). 423 */ 424int 425bwsfwrite(struct bwstring *bws, FILE *f, bool zero_ended) 426{ 427 428 if (MB_CUR_MAX == 1) { 429 size_t len = bws->len; 430 431 if (!zero_ended) { 432 bws->data.cstr[len] = '\n'; 433 434 if (fwrite(bws->data.cstr, len + 1, 1, f) < 1) 435 err(2, NULL); 436 437 bws->data.cstr[len] = '\0'; 438 } else if (fwrite(bws->data.cstr, len + 1, 1, f) < 1) 439 err(2, NULL); 440 441 return (len + 1); 442 443 } else { 444 wchar_t eols; 445 int printed = 0; 446 447 eols = zero_ended ? btowc('\0') : btowc('\n'); 448 449 while (printed < (int) BWSLEN(bws)) { 450 const wchar_t *s = bws->data.wstr + printed; 451 452 if (*s == L'\0') { 453 int nums; 454 455 nums = fwprintf(f, L"%lc", *s); 456 457 if (nums != 1) 458 err(2, NULL); 459 ++printed; 460 } else { 461 int nums; 462 463 nums = fwprintf(f, L"%ls", s); 464 465 if (nums < 1) 466 err(2, NULL); 467 printed += nums; 468 } 469 } 470 fwprintf(f, L"%lc", eols); 471 return (printed + 1); 472 } 473} 474 475/* 476 * Allocate and read a binary string from file. 477 * The strings are nl-ended or zero-ended, depending on the sort setting. 478 */ 479struct bwstring * 480bwsfgetln(FILE *f, size_t *len, bool zero_ended, struct reader_buffer *rb) 481{ 482 wchar_t eols; 483 484 eols = zero_ended ? btowc('\0') : btowc('\n'); 485 486 if (!zero_ended && (MB_CUR_MAX > 1)) { 487 wchar_t *ret; 488 489 ret = fgetwln(f, len); 490 491 if (ret == NULL) { 492 if (!feof(f)) 493 err(2, NULL); 494 return (NULL); 495 } 496 if (*len > 0) { 497 if (ret[*len - 1] == eols) 498 --(*len); 499 } 500 return (bwssbdup(ret, *len)); 501 502 } else if (!zero_ended && (MB_CUR_MAX == 1)) { 503 char *ret; 504 505 ret = fgetln(f, len); 506 507 if (ret == NULL) { 508 if (!feof(f)) 509 err(2, NULL); 510 return (NULL); 511 } 512 if (*len > 0) { 513 if (ret[*len - 1] == '\n') 514 --(*len); 515 } 516 return (bwscsbdup(ret, *len)); 517 518 } else { 519 wchar_t c = 0; 520 521 *len = 0; 522 523 if (feof(f)) 524 return (NULL); 525 526 if (2 >= rb->fgetwln_z_buffer_size) { 527 rb->fgetwln_z_buffer_size += 256; 528 rb->fgetwln_z_buffer = sort_realloc(rb->fgetwln_z_buffer, 529 sizeof(wchar_t) * rb->fgetwln_z_buffer_size); 530 } 531 rb->fgetwln_z_buffer[*len] = 0; 532 533 if (MB_CUR_MAX == 1) 534 while (!feof(f)) { 535 c = fgetc(f); 536 537 if (c == EOF) { 538 if (*len == 0) 539 return (NULL); 540 goto line_read_done; 541 } 542 if (c == eols) 543 goto line_read_done; 544 545 if (*len + 1 >= rb->fgetwln_z_buffer_size) { 546 rb->fgetwln_z_buffer_size += 256; 547 rb->fgetwln_z_buffer = sort_realloc(rb->fgetwln_z_buffer, 548 SIZEOF_WCHAR_STRING(rb->fgetwln_z_buffer_size)); 549 } 550 551 rb->fgetwln_z_buffer[*len] = c; 552 rb->fgetwln_z_buffer[++(*len)] = 0; 553 } 554 else 555 while (!feof(f)) { 556 c = fgetwc(f); 557 558 if (c == WEOF) { 559 if (*len == 0) 560 return (NULL); 561 goto line_read_done; 562 } 563 if (c == eols) 564 goto line_read_done; 565 566 if (*len + 1 >= rb->fgetwln_z_buffer_size) { 567 rb->fgetwln_z_buffer_size += 256; 568 rb->fgetwln_z_buffer = sort_realloc(rb->fgetwln_z_buffer, 569 SIZEOF_WCHAR_STRING(rb->fgetwln_z_buffer_size)); 570 } 571 572 rb->fgetwln_z_buffer[*len] = c; 573 rb->fgetwln_z_buffer[++(*len)] = 0; 574 } 575 576line_read_done: 577 /* we do not count the last 0 */ 578 return (bwssbdup(rb->fgetwln_z_buffer, *len)); 579 } 580} 581 582int 583bwsncmp(const struct bwstring *bws1, const struct bwstring *bws2, 584 size_t offset, size_t len) 585{ 586 size_t cmp_len, len1, len2; 587 int res = 0; 588 589 cmp_len = 0; 590 len1 = bws1->len; 591 len2 = bws2->len; 592 593 if (len1 <= offset) { 594 return ((len2 <= offset) ? 0 : -1); 595 } else { 596 if (len2 <= offset) 597 return (+1); 598 else { 599 len1 -= offset; 600 len2 -= offset; 601 602 cmp_len = len1; 603 604 if (len2 < cmp_len) 605 cmp_len = len2; 606 607 if (len < cmp_len) 608 cmp_len = len; 609 610 if (MB_CUR_MAX == 1) { 611 const unsigned char *s1, *s2; 612 613 s1 = bws1->data.cstr + offset; 614 s2 = bws2->data.cstr + offset; 615 616 res = memcmp(s1, s2, cmp_len); 617 618 } else { 619 const wchar_t *s1, *s2; 620 621 s1 = bws1->data.wstr + offset; 622 s2 = bws2->data.wstr + offset; 623 624 res = memcmp(s1, s2, SIZEOF_WCHAR_STRING(cmp_len)); 625 } 626 } 627 } 628 629 if (res == 0) { 630 if (len1 < cmp_len && len1 < len2) 631 res = -1; 632 else if (len2 < cmp_len && len2 < len1) 633 res = +1; 634 } 635 636 return (res); 637} 638 639int 640bwscmp(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset) 641{ 642 size_t len1, len2, cmp_len; 643 int res; 644 645 len1 = bws1->len; 646 len2 = bws2->len; 647 648 len1 -= offset; 649 len2 -= offset; 650 651 cmp_len = len1; 652 653 if (len2 < cmp_len) 654 cmp_len = len2; 655 656 res = bwsncmp(bws1, bws2, offset, cmp_len); 657 658 if (res == 0) { 659 if( len1 < len2) 660 res = -1; 661 else if (len2 < len1) 662 res = +1; 663 } 664 665 return (res); 666} 667 668int 669bws_iterator_cmp(bwstring_iterator iter1, bwstring_iterator iter2, size_t len) 670{ 671 wchar_t c1, c2; 672 size_t i = 0; 673 674 for (i = 0; i < len; ++i) { 675 c1 = bws_get_iter_value(iter1); 676 c2 = bws_get_iter_value(iter2); 677 if (c1 != c2) 678 return (c1 - c2); 679 iter1 = bws_iterator_inc(iter1, 1); 680 iter2 = bws_iterator_inc(iter2, 1); 681 } 682 683 return (0); 684} 685 686int 687bwscoll(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset) 688{ 689 size_t len1, len2; 690 691 len1 = bws1->len; 692 len2 = bws2->len; 693 694 if (len1 <= offset) 695 return ((len2 <= offset) ? 0 : -1); 696 else { 697 if (len2 <= offset) 698 return (+1); 699 else { 700 len1 -= offset; 701 len2 -= offset; 702 703 if (MB_CUR_MAX == 1) { 704 const unsigned char *s1, *s2; 705 706 s1 = bws1->data.cstr + offset; 707 s2 = bws2->data.cstr + offset; 708 709 if (byte_sort) { 710 int res = 0; 711 712 if (len1 > len2) { 713 res = memcmp(s1, s2, len2); 714 if (!res) 715 res = +1; 716 } else if (len1 < len2) { 717 res = memcmp(s1, s2, len1); 718 if (!res) 719 res = -1; 720 } else 721 res = memcmp(s1, s2, len1); 722 723 return (res); 724 725 } else { 726 int res = 0; 727 size_t i, maxlen; 728 729 i = 0; 730 maxlen = len1; 731 732 if (maxlen > len2) 733 maxlen = len2; 734 735 while (i < maxlen) { 736 /* goto next non-zero part: */ 737 while ((i < maxlen) && 738 !s1[i] && !s2[i]) 739 ++i; 740 741 if (i >= maxlen) 742 break; 743 744 if (s1[i] == 0) { 745 if (s2[i] == 0) 746 /* NOTREACHED */ 747 err(2, "bwscoll error 01"); 748 else 749 return (-1); 750 } else if (s2[i] == 0) 751 return (+1); 752 753 res = strcoll(s1 + i, s2 + i); 754 if (res) 755 return (res); 756 757 while ((i < maxlen) && 758 s1[i] && s2[i]) 759 ++i; 760 761 if (i >= maxlen) 762 break; 763 764 if (s1[i] == 0) { 765 if (s2[i] == 0) { 766 ++i; 767 continue; 768 } else 769 return (-1); 770 } else if (s2[i] == 0) 771 return (+1); 772 else 773 /* NOTREACHED */ 774 err(2, "bwscoll error 02"); 775 } 776 777 if (len1 < len2) 778 return (-1); 779 else if (len1 > len2) 780 return (+1); 781 782 return (0); 783 } 784 } else { 785 const wchar_t *s1, *s2; 786 size_t i, maxlen; 787 int res = 0; 788 789 s1 = bws1->data.wstr + offset; 790 s2 = bws2->data.wstr + offset; 791 792 i = 0; 793 maxlen = len1; 794 795 if (maxlen > len2) 796 maxlen = len2; 797 798 while (i < maxlen) { 799 800 /* goto next non-zero part: */ 801 while ((i < maxlen) && 802 !s1[i] && !s2[i]) 803 ++i; 804 805 if (i >= maxlen) 806 break; 807 808 if (s1[i] == 0) { 809 if (s2[i] == 0) 810 /* NOTREACHED */ 811 err(2, "bwscoll error 1"); 812 else 813 return (-1); 814 } else if (s2[i] == 0) 815 return (+1); 816 817 res = wide_str_coll(s1 + i, s2 + i); 818 if (res) 819 return (res); 820 821 while ((i < maxlen) && s1[i] && s2[i]) 822 ++i; 823 824 if (i >= maxlen) 825 break; 826 827 if (s1[i] == 0) { 828 if (s2[i] == 0) { 829 ++i; 830 continue; 831 } else 832 return (-1); 833 } else if (s2[i] == 0) 834 return (+1); 835 else 836 /* NOTREACHED */ 837 err(2, "bwscoll error 2"); 838 } 839 840 if (len1 < len2) 841 return (-1); 842 else if (len1 > len2) 843 return (+1); 844 845 return (0); 846 } 847 } 848 } 849} 850 851/* 852 * Correction of the system API 853 */ 854double 855bwstod(struct bwstring *s0, bool *empty) 856{ 857 double ret = 0; 858 859 if (MB_CUR_MAX == 1) { 860 unsigned char *end, *s; 861 char *ep; 862 863 s = s0->data.cstr; 864 end = s + s0->len; 865 ep = NULL; 866 867 while (isblank(*s) && s < end) 868 ++s; 869 870 if (!isprint(*s)) { 871 *empty = true; 872 return (0); 873 } 874 875 ret = strtod(s, &ep); 876 if ((unsigned char*) ep == s) { 877 *empty = true; 878 return (0); 879 } 880 } else { 881 wchar_t *end, *ep, *s; 882 883 s = s0->data.wstr; 884 end = s + s0->len; 885 ep = NULL; 886 887 while (iswblank(*s) && s < end) 888 ++s; 889 890 if (!iswprint(*s)) { 891 *empty = true; 892 return (0); 893 } 894 895 ret = wcstod(s, &ep); 896 if (ep == s) { 897 *empty = true; 898 return (0); 899 } 900 } 901 902 *empty = false; 903 return (ret); 904} 905 906/* 907 * A helper function for monthcoll. If a line matches 908 * a month name, it returns (number of the month - 1), 909 * while if there is no match, it just return -1. 910 */ 911 912int 913bws_month_score(const struct bwstring *s0) 914{ 915 916 if (MB_CUR_MAX == 1) { 917 const unsigned char *end, *s; 918 size_t len; 919 920 s = s0->data.cstr; 921 end = s + s0->len; 922 923 while (isblank(*s) && s < end) 924 ++s; 925 926 len = strlen(s); 927 928 for (int i = 11; i >= 0; --i) { 929 if (cmonths[i] && 930 (s == (unsigned char*)strstr(s, cmonths[i]))) 931 return (i); 932 } 933 934 } else { 935 const wchar_t *end, *s; 936 size_t len; 937 938 s = s0->data.wstr; 939 end = s + s0->len; 940 941 while (iswblank(*s) && s < end) 942 ++s; 943 944 len = wcslen(s); 945 946 for (int i = 11; i >= 0; --i) { 947 if (wmonths[i] && (s == wcsstr(s, wmonths[i]))) 948 return (i); 949 } 950 } 951 952 return (-1); 953} 954 955/* 956 * Rips out leading blanks (-b). 957 */ 958struct bwstring * 959ignore_leading_blanks(struct bwstring *str) 960{ 961 962 if (MB_CUR_MAX == 1) { 963 unsigned char *dst, *end, *src; 964 965 src = str->data.cstr; 966 dst = src; 967 end = src + str->len; 968 969 while (src < end && isblank(*src)) 970 ++src; 971 972 if (src != dst) { 973 size_t newlen; 974 975 newlen = BWSLEN(str) - (src - dst); 976 977 while (src < end) { 978 *dst = *src; 979 ++dst; 980 ++src; 981 } 982 bws_setlen(str, newlen); 983 } 984 } else { 985 wchar_t *dst, *end, *src; 986 987 src = str->data.wstr; 988 dst = src; 989 end = src + str->len; 990 991 while (src < end && iswblank(*src)) 992 ++src; 993 994 if (src != dst) { 995 996 size_t newlen = BWSLEN(str) - (src - dst); 997 998 while (src < end) { 999 *dst = *src; 1000 ++dst; 1001 ++src; 1002 } 1003 bws_setlen(str, newlen); 1004 1005 } 1006 } 1007 return (str); 1008} 1009 1010/* 1011 * Rips out nonprinting characters (-i). 1012 */ 1013struct bwstring * 1014ignore_nonprinting(struct bwstring *str) 1015{ 1016 size_t newlen = str->len; 1017 1018 if (MB_CUR_MAX == 1) { 1019 unsigned char *dst, *end, *src; 1020 unsigned char c; 1021 1022 src = str->data.cstr; 1023 dst = src; 1024 end = src + str->len; 1025 1026 while (src < end) { 1027 c = *src; 1028 if (isprint(c)) { 1029 *dst = c; 1030 ++dst; 1031 ++src; 1032 } else { 1033 ++src; 1034 --newlen; 1035 } 1036 } 1037 } else { 1038 wchar_t *dst, *end, *src; 1039 wchar_t c; 1040 1041 src = str->data.wstr; 1042 dst = src; 1043 end = src + str->len; 1044 1045 while (src < end) { 1046 c = *src; 1047 if (iswprint(c)) { 1048 *dst = c; 1049 ++dst; 1050 ++src; 1051 } else { 1052 ++src; 1053 --newlen; 1054 } 1055 } 1056 } 1057 bws_setlen(str, newlen); 1058 1059 return (str); 1060} 1061 1062/* 1063 * Rips out any characters that are not alphanumeric characters 1064 * nor blanks (-d). 1065 */ 1066struct bwstring * 1067dictionary_order(struct bwstring *str) 1068{ 1069 size_t newlen = str->len; 1070 1071 if (MB_CUR_MAX == 1) { 1072 unsigned char *dst, *end, *src; 1073 unsigned char c; 1074 1075 src = str->data.cstr; 1076 dst = src; 1077 end = src + str->len; 1078 1079 while (src < end) { 1080 c = *src; 1081 if (isalnum(c) || isblank(c)) { 1082 *dst = c; 1083 ++dst; 1084 ++src; 1085 } else { 1086 ++src; 1087 --newlen; 1088 } 1089 } 1090 } else { 1091 wchar_t *dst, *end, *src; 1092 wchar_t c; 1093 1094 src = str->data.wstr; 1095 dst = src; 1096 end = src + str->len; 1097 1098 while (src < end) { 1099 c = *src; 1100 if (iswalnum(c) || iswblank(c)) { 1101 *dst = c; 1102 ++dst; 1103 ++src; 1104 } else { 1105 ++src; 1106 --newlen; 1107 } 1108 } 1109 } 1110 bws_setlen(str, newlen); 1111 1112 return (str); 1113} 1114 1115/* 1116 * Converts string to lower case(-f). 1117 */ 1118struct bwstring * 1119ignore_case(struct bwstring *str) 1120{ 1121 1122 if (MB_CUR_MAX == 1) { 1123 unsigned char *end, *s; 1124 1125 s = str->data.cstr; 1126 end = s + str->len; 1127 1128 while (s < end) { 1129 *s = toupper(*s); 1130 ++s; 1131 } 1132 } else { 1133 wchar_t *end, *s; 1134 1135 s = str->data.wstr; 1136 end = s + str->len; 1137 1138 while (s < end) { 1139 *s = towupper(*s); 1140 ++s; 1141 } 1142 } 1143 return (str); 1144} 1145 1146void 1147bws_disorder_warnx(struct bwstring *s, const char *fn, size_t pos) 1148{ 1149 1150 if (MB_CUR_MAX == 1) 1151 warnx("%s:%zu: disorder: %s", fn, pos + 1, s->data.cstr); 1152 else 1153 warnx("%s:%zu: disorder: %ls", fn, pos + 1, s->data.wstr); 1154} 1155