Annotation of src/usr.bin/sort/bwstring.c, Revision 1.9
1.9 ! schwarze 1: /* $OpenBSD: bwstring.c,v 1.8 2019/05/15 09:07:46 schwarze Exp $ */
1.1 millert 2:
3: /*-
4: * Copyright (C) 2009 Gabor Kovesdan <gabor@FreeBSD.org>
5: * Copyright (C) 2012 Oleg Moskalenko <mom040267@gmail.com>
6: * All rights reserved.
7: *
8: * Redistribution and use in source and binary forms, with or without
9: * modification, are permitted provided that the following conditions
10: * are met:
11: * 1. Redistributions of source code must retain the above copyright
12: * notice, this list of conditions and the following disclaimer.
13: * 2. Redistributions in binary form must reproduce the above copyright
14: * notice, this list of conditions and the following disclaimer in the
15: * documentation and/or other materials provided with the distribution.
16: *
17: * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20: * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27: * SUCH DAMAGE.
28: */
29:
30: #include <ctype.h>
31: #include <errno.h>
32: #include <err.h>
33: #include <langinfo.h>
34: #include <math.h>
35: #include <stdlib.h>
36: #include <string.h>
37: #include <wchar.h>
38: #include <wctype.h>
39:
40: #include "bwstring.h"
41: #include "sort.h"
42:
43: static wchar_t **wmonths;
44: static char **cmonths;
45:
46: /* initialise months */
47:
48: void
49: initialise_months(void)
50: {
51: const nl_item item[12] = { ABMON_1, ABMON_2, ABMON_3, ABMON_4,
52: ABMON_5, ABMON_6, ABMON_7, ABMON_8, ABMON_9, ABMON_10,
53: ABMON_11, ABMON_12 };
54: char *tmp;
55: size_t len;
56:
57: if (sort_mb_cur_max == 1) {
58: if (cmonths == NULL) {
59: char *m;
60: unsigned int j;
61: int i;
62:
63: cmonths = sort_malloc(sizeof(char *) * 12);
64: for (i = 0; i < 12; i++) {
65: cmonths[i] = NULL;
66: tmp = nl_langinfo(item[i]);
67: if (debug_sort)
68: printf("month[%d]=%s\n", i, tmp);
1.3 millert 69: if (*tmp == '\0')
1.1 millert 70: continue;
1.3 millert 71: m = sort_strdup(tmp);
1.4 millert 72: len = strlen(tmp);
1.1 millert 73: for (j = 0; j < len; j++)
74: m[j] = toupper(m[j]);
75: cmonths[i] = m;
76: }
77: }
78: } else {
79: if (wmonths == NULL) {
80: unsigned int j;
81: wchar_t *m;
82: int i;
83:
84: wmonths = sort_malloc(sizeof(wchar_t *) * 12);
85: for (i = 0; i < 12; i++) {
86: wmonths[i] = NULL;
87: tmp = nl_langinfo(item[i]);
88: if (debug_sort)
89: printf("month[%d]=%s\n", i, tmp);
1.3 millert 90: if (*tmp == '\0')
91: continue;
1.1 millert 92: len = strlen(tmp);
1.7 millert 93: m = sort_reallocarray(NULL, len + 1,
94: sizeof(wchar_t));
1.3 millert 95: if (mbstowcs(m, tmp, len) == (size_t)-1) {
96: sort_free(m);
1.1 millert 97: continue;
1.3 millert 98: }
1.1 millert 99: m[len] = L'\0';
100: for (j = 0; j < len; j++)
101: m[j] = towupper(m[j]);
102: wmonths[i] = m;
103: }
104: }
105: }
106: }
107:
108: /*
109: * Compare two wide-character strings
110: */
111: static int
112: wide_str_coll(const wchar_t *s1, const wchar_t *s2)
113: {
114: int ret = 0;
115:
116: errno = 0;
117: ret = wcscoll(s1, s2);
118: if (errno == EILSEQ) {
119: errno = 0;
120: ret = wcscmp(s1, s2);
121: if (errno != 0) {
122: size_t i;
123: for (i = 0; ; ++i) {
124: wchar_t c1 = s1[i];
125: wchar_t c2 = s2[i];
126: if (c1 == L'\0')
127: return (c2 == L'\0') ? 0 : -1;
128: if (c2 == L'\0')
129: return 1;
130: if (c1 == c2)
131: continue;
132: return (int)c1 - (int)c2;
133: }
134: }
135: }
136: return ret;
137: }
138:
139: /* counterparts of wcs functions */
140:
141: void
142: bwsprintf(FILE *f, struct bwstring *bws, const char *prefix, const char *suffix)
143: {
144: if (sort_mb_cur_max == 1)
145: fprintf(f, "%s%s%s", prefix, bws->data.cstr, suffix);
146: else
147: fprintf(f, "%s%S%s", prefix, bws->data.wstr, suffix);
148: }
149:
150: const void *
151: bwsrawdata(const struct bwstring *bws)
152: {
153: return &(bws->data);
154: }
155:
156: size_t
157: bwsrawlen(const struct bwstring *bws)
158: {
159: return (sort_mb_cur_max == 1) ? bws->len : SIZEOF_WCHAR_STRING(bws->len);
160: }
161:
162: size_t
163: bws_memsize(const struct bwstring *bws)
164: {
165: return (sort_mb_cur_max == 1) ? (bws->len + 2 + sizeof(struct bwstring)) :
166: (SIZEOF_WCHAR_STRING(bws->len + 1) + sizeof(struct bwstring));
167: }
168:
169: void
170: bws_setlen(struct bwstring *bws, size_t newlen)
171: {
172: if (bws && newlen != bws->len && newlen <= bws->len) {
173: bws->len = newlen;
174: if (sort_mb_cur_max == 1)
175: bws->data.cstr[newlen] = '\0';
176: else
177: bws->data.wstr[newlen] = L'\0';
178: }
179: }
180:
181: /*
182: * Allocate a new binary string of specified size
183: */
184: struct bwstring *
185: bwsalloc(size_t sz)
186: {
187: struct bwstring *ret;
188:
189: if (sort_mb_cur_max == 1) {
190: ret = sort_malloc(sizeof(struct bwstring) + 1 + sz);
191: ret->data.cstr[sz] = '\0';
192: } else {
193: ret = sort_malloc(sizeof(struct bwstring) +
194: SIZEOF_WCHAR_STRING(sz + 1));
195: ret->data.wstr[sz] = L'\0';
196: }
197: ret->len = sz;
198:
199: return ret;
200: }
201:
202: /*
203: * Create a copy of binary string.
204: * New string size equals the length of the old string.
205: */
206: struct bwstring *
207: bwsdup(const struct bwstring *s)
208: {
209: struct bwstring *ret;
210:
211: if (s == NULL)
212: return NULL;
213:
214: ret = bwsalloc(s->len);
215:
216: if (sort_mb_cur_max == 1)
217: memcpy(ret->data.cstr, s->data.cstr, s->len);
218: else
219: memcpy(ret->data.wstr, s->data.wstr,
220: SIZEOF_WCHAR_STRING(s->len));
221:
222: return ret;
223: }
224:
225: /*
1.5 millert 226: * Create a new binary string from a wide character buffer.
1.1 millert 227: */
228: struct bwstring *
229: bwssbdup(const wchar_t *str, size_t len)
230: {
231: if (str == NULL)
232: return (len == 0) ? bwsalloc(0) : NULL;
233: else {
234: struct bwstring *ret;
235: size_t i;
236:
237: ret = bwsalloc(len);
238:
239: if (sort_mb_cur_max == 1)
240: for (i = 0; i < len; ++i)
241: ret->data.cstr[i] = (unsigned char) str[i];
242: else
243: memcpy(ret->data.wstr, str, SIZEOF_WCHAR_STRING(len));
244:
245: return ret;
246: }
247: }
248:
249: /*
250: * Create a new binary string from a raw binary buffer.
251: */
252: struct bwstring *
253: bwscsbdup(const unsigned char *str, size_t len)
254: {
255: struct bwstring *ret;
256:
257: ret = bwsalloc(len);
258:
259: if (str) {
260: if (sort_mb_cur_max == 1)
261: memcpy(ret->data.cstr, str, len);
262: else {
263: mbstate_t mbs;
264: const char *s;
265: size_t charlen, chars, cptr;
266:
267: chars = 0;
268: cptr = 0;
269: s = (const char *) str;
270:
271: memset(&mbs, 0, sizeof(mbs));
272:
273: while (cptr < len) {
274: size_t n = sort_mb_cur_max;
275:
276: if (n > len - cptr)
277: n = len - cptr;
278: charlen = mbrlen(s + cptr, n, &mbs);
279: switch (charlen) {
280: case 0:
281: /* FALLTHROUGH */
282: case (size_t) -1:
283: /* FALLTHROUGH */
284: case (size_t) -2:
285: ret->data.wstr[chars++] =
286: (unsigned char) s[cptr];
287: ++cptr;
288: break;
289: default:
290: n = mbrtowc(ret->data.wstr + (chars++),
291: s + cptr, charlen, &mbs);
292: if ((n == (size_t)-1) || (n == (size_t)-2))
293: /* NOTREACHED */
294: err(2, "mbrtowc error");
295: cptr += charlen;
296: };
297: }
298:
299: ret->len = chars;
300: ret->data.wstr[ret->len] = L'\0';
301: }
302: }
303: return ret;
304: }
305:
306: /*
307: * De-allocate object memory
308: */
309: void
310: bwsfree(struct bwstring *s)
311: {
312: sort_free(s);
313: }
314:
315: /*
316: * Copy content of src binary string to dst.
317: * If the capacity of the dst string is not sufficient,
318: * then the data is truncated.
319: */
320: size_t
321: bwscpy(struct bwstring *dst, const struct bwstring *src)
322: {
323: size_t nums = src->len;
324:
325: if (nums > dst->len)
326: nums = dst->len;
327: dst->len = nums;
328:
329: if (sort_mb_cur_max == 1) {
330: memcpy(dst->data.cstr, src->data.cstr, nums);
331: dst->data.cstr[dst->len] = '\0';
332: } else {
333: memcpy(dst->data.wstr, src->data.wstr,
334: SIZEOF_WCHAR_STRING(nums + 1));
335: dst->data.wstr[dst->len] = L'\0';
336: }
337:
338: return nums;
339: }
340:
341: /*
342: * Copy content of src binary string to dst,
343: * with specified number of symbols to be copied.
344: * If the capacity of the dst string is not sufficient,
345: * then the data is truncated.
346: */
347: struct bwstring *
348: bwsncpy(struct bwstring *dst, const struct bwstring *src, size_t size)
349: {
350: size_t nums = src->len;
351:
352: if (nums > dst->len)
353: nums = dst->len;
354: if (nums > size)
355: nums = size;
356: dst->len = nums;
357:
358: if (sort_mb_cur_max == 1) {
359: memcpy(dst->data.cstr, src->data.cstr, nums);
360: dst->data.cstr[dst->len] = '\0';
361: } else {
362: memcpy(dst->data.wstr, src->data.wstr,
363: SIZEOF_WCHAR_STRING(nums + 1));
364: dst->data.wstr[dst->len] = L'\0';
365: }
366:
367: return dst;
368: }
369:
370: /*
371: * Copy content of src binary string to dst,
372: * with specified number of symbols to be copied.
373: * An offset value can be specified, from the start of src string.
374: * If the capacity of the dst string is not sufficient,
375: * then the data is truncated.
376: */
377: struct bwstring *
378: bwsnocpy(struct bwstring *dst, const struct bwstring *src, size_t offset,
379: size_t size)
380: {
381: if (offset >= src->len) {
382: dst->data.wstr[0] = 0;
383: dst->len = 0;
384: } else {
385: size_t nums = src->len - offset;
386:
387: if (nums > dst->len)
388: nums = dst->len;
389: if (nums > size)
390: nums = size;
391: dst->len = nums;
392: if (sort_mb_cur_max == 1) {
393: memcpy(dst->data.cstr, src->data.cstr + offset,
394: (nums));
395: dst->data.cstr[dst->len] = '\0';
396: } else {
397: memcpy(dst->data.wstr, src->data.wstr + offset,
398: SIZEOF_WCHAR_STRING(nums));
399: dst->data.wstr[dst->len] = L'\0';
400: }
401: }
402: return dst;
403: }
404:
405: /*
406: * Write binary string to the file.
407: * The output is ended either with '\n' (nl == true)
408: * or '\0' (nl == false).
409: */
410: size_t
411: bwsfwrite(struct bwstring *bws, FILE *f, bool zero_ended)
412: {
413: if (sort_mb_cur_max == 1) {
414: size_t len = bws->len;
415:
416: if (!zero_ended) {
417: bws->data.cstr[len] = '\n';
418:
419: if (fwrite(bws->data.cstr, len + 1, 1, f) < 1)
420: err(2, NULL);
421:
422: bws->data.cstr[len] = '\0';
423: } else if (fwrite(bws->data.cstr, len + 1, 1, f) < 1)
424: err(2, NULL);
425:
426: return len + 1;
427:
428: } else {
429: wchar_t eols;
430: size_t printed = 0;
431:
432: eols = zero_ended ? btowc('\0') : btowc('\n');
433:
434: while (printed < BWSLEN(bws)) {
435: const wchar_t *s = bws->data.wstr + printed;
436:
437: if (*s == L'\0') {
438: int nums;
439:
440: nums = fwprintf(f, L"%lc", *s);
441:
442: if (nums != 1)
443: err(2, NULL);
444: ++printed;
445: } else {
446: int nums;
447:
448: nums = fwprintf(f, L"%ls", s);
449:
450: if (nums < 1)
451: err(2, NULL);
452: printed += nums;
453: }
454: }
455: fwprintf(f, L"%lc", eols);
456: return printed + 1;
457: }
458: }
459:
460: /*
461: * Allocate and read a binary string from file.
462: * The strings are nl-ended or zero-ended, depending on the sort setting.
463: */
464: struct bwstring *
465: bwsfgetln(FILE *f, size_t *len, bool zero_ended, struct reader_buffer *rb)
466: {
467: wint_t eols;
468:
469: eols = zero_ended ? btowc('\0') : btowc('\n');
470:
471: if (!zero_ended && (sort_mb_cur_max > 1)) {
472: wchar_t *ret;
473:
474: ret = fgetwln(f, len);
475:
476: if (ret == NULL) {
477: if (!feof(f))
478: err(2, NULL);
479: return NULL;
480: }
481: if (*len > 0) {
482: if (ret[*len - 1] == (wchar_t)eols)
483: --(*len);
484: }
485: return bwssbdup(ret, *len);
486:
487: } else if (!zero_ended && (sort_mb_cur_max == 1)) {
488: char *ret;
489:
490: ret = fgetln(f, len);
491:
492: if (ret == NULL) {
493: if (!feof(f))
494: err(2, NULL);
495: return NULL;
496: }
497: if (*len > 0) {
498: if (ret[*len - 1] == '\n')
499: --(*len);
500: }
501: return bwscsbdup((unsigned char *)ret, *len);
502:
503: } else {
504: *len = 0;
505:
506: if (feof(f))
507: return NULL;
508:
509: if (2 >= rb->fgetwln_z_buffer_size) {
510: rb->fgetwln_z_buffer_size += 256;
511: rb->fgetwln_z_buffer =
512: sort_reallocarray(rb->fgetwln_z_buffer,
513: rb->fgetwln_z_buffer_size, sizeof(wchar_t));
514: }
515: rb->fgetwln_z_buffer[*len] = 0;
516:
1.6 millert 517: if (sort_mb_cur_max == 1) {
1.1 millert 518: while (!feof(f)) {
519: int c;
520:
521: c = fgetc(f);
522:
523: if (c == EOF) {
524: if (*len == 0)
525: return NULL;
526: goto line_read_done;
527: }
528: if (c == eols)
529: goto line_read_done;
530:
531: if (*len + 1 >= rb->fgetwln_z_buffer_size) {
532: rb->fgetwln_z_buffer_size += 256;
533: rb->fgetwln_z_buffer =
534: sort_reallocarray(rb->fgetwln_z_buffer,
535: rb->fgetwln_z_buffer_size, sizeof(wchar_t));
536: }
537:
538: rb->fgetwln_z_buffer[*len] = c;
539: rb->fgetwln_z_buffer[++(*len)] = 0;
540: }
1.6 millert 541: } else {
1.1 millert 542: while (!feof(f)) {
543: wint_t c = 0;
544:
545: c = fgetwc(f);
546:
547: if (c == WEOF) {
548: if (*len == 0)
549: return NULL;
550: goto line_read_done;
551: }
552: if (c == eols)
553: goto line_read_done;
554:
555: if (*len + 1 >= rb->fgetwln_z_buffer_size) {
556: rb->fgetwln_z_buffer_size += 256;
557: rb->fgetwln_z_buffer =
558: sort_reallocarray(rb->fgetwln_z_buffer,
559: rb->fgetwln_z_buffer_size, sizeof(wchar_t));
560: }
561:
562: rb->fgetwln_z_buffer[*len] = c;
563: rb->fgetwln_z_buffer[++(*len)] = 0;
564: }
1.6 millert 565: }
1.1 millert 566:
567: line_read_done:
568: /* we do not count the last 0 */
569: return bwssbdup(rb->fgetwln_z_buffer, *len);
570: }
571: }
572:
573: int
574: bwsncmp(const struct bwstring *bws1, const struct bwstring *bws2,
575: size_t offset, size_t len)
576: {
577: size_t cmp_len, len1, len2;
578: int res = 0;
579:
580: len1 = bws1->len;
581: len2 = bws2->len;
582:
583: if (len1 <= offset) {
584: return (len2 <= offset) ? 0 : -1;
585: } else {
586: if (len2 <= offset)
587: return 1;
588: else {
589: len1 -= offset;
590: len2 -= offset;
591:
592: cmp_len = len1;
593:
594: if (len2 < cmp_len)
595: cmp_len = len2;
596:
597: if (len < cmp_len)
598: cmp_len = len;
599:
600: if (sort_mb_cur_max == 1) {
601: const unsigned char *s1, *s2;
602:
603: s1 = bws1->data.cstr + offset;
604: s2 = bws2->data.cstr + offset;
605:
606: res = memcmp(s1, s2, cmp_len);
607:
608: } else {
609: const wchar_t *s1, *s2;
610:
611: s1 = bws1->data.wstr + offset;
612: s2 = bws2->data.wstr + offset;
613:
614: res = memcmp(s1, s2, SIZEOF_WCHAR_STRING(cmp_len));
615: }
616: }
617: }
618:
619: if (res == 0) {
620: if (len1 < cmp_len && len1 < len2)
621: res = -1;
622: else if (len2 < cmp_len && len2 < len1)
623: res = +1;
624: }
625:
626: return res;
627: }
628:
629: int
630: bwscmp(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset)
631: {
632: size_t len1, len2, cmp_len;
633: int res;
634:
635: len1 = bws1->len;
636: len2 = bws2->len;
637:
638: len1 -= offset;
639: len2 -= offset;
640:
641: cmp_len = len1;
642:
643: if (len2 < cmp_len)
644: cmp_len = len2;
645:
646: res = bwsncmp(bws1, bws2, offset, cmp_len);
647:
648: if (res == 0) {
649: if (len1 < len2)
650: res = -1;
651: else if (len2 < len1)
652: res = +1;
653: }
654:
655: return res;
656: }
657:
658: int
659: bws_iterator_cmp(bwstring_iterator iter1, bwstring_iterator iter2, size_t len)
660: {
661: wchar_t c1, c2;
662: size_t i = 0;
663:
664: for (i = 0; i < len; ++i) {
665: c1 = bws_get_iter_value(iter1);
666: c2 = bws_get_iter_value(iter2);
667: if (c1 != c2)
668: return c1 - c2;
669: iter1 = bws_iterator_inc(iter1, 1);
670: iter2 = bws_iterator_inc(iter2, 1);
671: }
672:
673: return 0;
674: }
675:
676: int
677: bwscoll(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset)
678: {
679: size_t len1, len2;
680:
681: len1 = bws1->len;
682: len2 = bws2->len;
683:
684: if (len1 <= offset)
685: return (len2 <= offset) ? 0 : -1;
1.8 schwarze 686:
1.9 ! schwarze 687: if (len2 <= offset)
! 688: return 1;
1.8 schwarze 689:
1.9 ! schwarze 690: len1 -= offset;
! 691: len2 -= offset;
1.1 millert 692:
1.9 ! schwarze 693: if (sort_mb_cur_max == 1) {
! 694: const unsigned char *s1, *s2;
! 695: int res;
1.1 millert 696:
1.9 ! schwarze 697: s1 = bws1->data.cstr + offset;
! 698: s2 = bws2->data.cstr + offset;
1.1 millert 699:
1.9 ! schwarze 700: if (len1 > len2) {
! 701: res = memcmp(s1, s2, len2);
! 702: if (!res)
! 703: res = +1;
! 704: } else if (len1 < len2) {
! 705: res = memcmp(s1, s2, len1);
! 706: if (!res)
! 707: res = -1;
! 708: } else
! 709: res = memcmp(s1, s2, len1);
1.1 millert 710:
1.9 ! schwarze 711: return res;
! 712: } else {
! 713: const wchar_t *s1, *s2;
! 714: size_t i, maxlen;
! 715: int res = 0;
! 716:
! 717: s1 = bws1->data.wstr + offset;
! 718: s2 = bws2->data.wstr + offset;
! 719:
! 720: i = 0;
! 721: maxlen = len1;
! 722:
! 723: if (maxlen > len2)
! 724: maxlen = len2;
! 725:
! 726: while (i < maxlen) {
! 727:
! 728: /* goto next non-zero part: */
! 729: while (i < maxlen &&
! 730: s1[i] == L'\0' && s2[i] == L'\0')
! 731: ++i;
! 732:
! 733: if (i >= maxlen)
! 734: break;
! 735:
! 736: if (s1[i] == L'\0') {
! 737: if (s2[i] == L'\0')
! 738: /* NOTREACHED */
! 739: err(2, "bwscoll error 1");
! 740: else
! 741: return -1;
! 742: } else if (s2[i] == L'\0')
! 743: return 1;
! 744:
! 745: res = wide_str_coll(s1 + i, s2 + i);
! 746: if (res)
! 747: return res;
! 748:
! 749: while (i < maxlen && s1[i] != L'\0' && s2[i] != L'\0')
! 750: ++i;
! 751:
! 752: if (i >= maxlen)
! 753: break;
! 754:
! 755: if (s1[i] == L'\0') {
! 756: if (s2[i] == L'\0') {
! 757: ++i;
! 758: continue;
! 759: } else
! 760: return -1;
! 761: } else if (s2[i] == L'\0')
! 762: return 1;
! 763: else
! 764: /* NOTREACHED */
! 765: err(2, "bwscoll error 2");
! 766: }
1.1 millert 767:
1.9 ! schwarze 768: if (len1 == len2)
! 769: return 0;
! 770: return len1 < len2 ? -1 : 1;
! 771: }
1.1 millert 772: }
773:
774: /*
775: * Correction of the system API
776: */
777: double
778: bwstod(struct bwstring *s0, bool *empty)
779: {
780: double ret = 0;
781:
782: if (sort_mb_cur_max == 1) {
783: char *ep, *end, *s;
784:
785: s = (char *)s0->data.cstr;
786: end = s + s0->len;
787: ep = NULL;
788:
789: while (isblank((unsigned char)*s) && s < end)
790: ++s;
791:
792: if (!isprint((unsigned char)*s)) {
793: *empty = true;
794: return 0;
795: }
796:
797: ret = strtod(s, &ep);
798: if (ep == s) {
799: *empty = true;
800: return 0;
801: }
802: } else {
803: wchar_t *end, *ep, *s;
804:
805: s = s0->data.wstr;
806: end = s + s0->len;
807: ep = NULL;
808:
809: while (iswblank(*s) && s < end)
810: ++s;
811:
812: if (!iswprint(*s)) {
813: *empty = true;
814: return 0;
815: }
816:
817: ret = wcstod(s, &ep);
818: if (ep == s) {
819: *empty = true;
820: return 0;
821: }
822: }
823:
824: *empty = false;
825: return ret;
826: }
827:
828: /*
829: * A helper function for monthcoll. If a line matches
830: * a month name, it returns (number of the month - 1),
831: * while if there is no match, it just return -1.
832: */
833: int
834: bws_month_score(const struct bwstring *s0)
835: {
836: if (sort_mb_cur_max == 1) {
837: const char *end, *s;
838: int i;
839:
840: s = (char *)s0->data.cstr;
841: end = s + s0->len;
842:
843: while (isblank((unsigned char)*s) && s < end)
844: ++s;
845:
846: for (i = 11; i >= 0; --i) {
847: if (cmonths[i] &&
848: (s == strstr(s, cmonths[i])))
849: return i;
850: }
851: } else {
852: const wchar_t *end, *s;
853: int i;
854:
855: s = s0->data.wstr;
856: end = s + s0->len;
857:
858: while (iswblank(*s) && s < end)
859: ++s;
860:
861: for (i = 11; i >= 0; --i) {
862: if (wmonths[i] && (s == wcsstr(s, wmonths[i])))
863: return i;
864: }
865: }
866:
867: return -1;
868: }
869:
870: /*
871: * Rips out leading blanks (-b).
872: */
873: struct bwstring *
874: ignore_leading_blanks(struct bwstring *str)
875: {
876: if (sort_mb_cur_max == 1) {
877: unsigned char *dst, *end, *src;
878:
879: src = str->data.cstr;
880: dst = src;
881: end = src + str->len;
882:
883: while (src < end && isblank(*src))
884: ++src;
885:
886: if (src != dst) {
887: size_t newlen;
888:
889: newlen = BWSLEN(str) - (src - dst);
890:
891: while (src < end) {
892: *dst = *src;
893: ++dst;
894: ++src;
895: }
896: bws_setlen(str, newlen);
897: }
898: } else {
899: wchar_t *dst, *end, *src;
900:
901: src = str->data.wstr;
902: dst = src;
903: end = src + str->len;
904:
905: while (src < end && iswblank(*src))
906: ++src;
907:
908: if (src != dst) {
909:
910: size_t newlen = BWSLEN(str) - (src - dst);
911:
912: while (src < end) {
913: *dst = *src;
914: ++dst;
915: ++src;
916: }
917: bws_setlen(str, newlen);
918:
919: }
920: }
921: return str;
922: }
923:
924: /*
925: * Rips out nonprinting characters (-i).
926: */
927: struct bwstring *
928: ignore_nonprinting(struct bwstring *str)
929: {
930: size_t newlen = str->len;
931:
932: if (sort_mb_cur_max == 1) {
933: unsigned char *dst, *end, *src;
934: unsigned char c;
935:
936: src = str->data.cstr;
937: dst = src;
938: end = src + str->len;
939:
940: while (src < end) {
941: c = *src;
942: if (isprint(c)) {
943: *dst = c;
944: ++dst;
945: ++src;
946: } else {
947: ++src;
948: --newlen;
949: }
950: }
951: } else {
952: wchar_t *dst, *end, *src;
953: wchar_t c;
954:
955: src = str->data.wstr;
956: dst = src;
957: end = src + str->len;
958:
959: while (src < end) {
960: c = *src;
961: if (iswprint(c)) {
962: *dst = c;
963: ++dst;
964: ++src;
965: } else {
966: ++src;
967: --newlen;
968: }
969: }
970: }
971: bws_setlen(str, newlen);
972:
973: return str;
974: }
975:
976: /*
977: * Rips out any characters that are not alphanumeric characters
978: * nor blanks (-d).
979: */
980: struct bwstring *
981: dictionary_order(struct bwstring *str)
982: {
983: size_t newlen = str->len;
984:
985: if (sort_mb_cur_max == 1) {
986: unsigned char *dst, *end, *src;
987: unsigned char c;
988:
989: src = str->data.cstr;
990: dst = src;
991: end = src + str->len;
992:
993: while (src < end) {
994: c = *src;
995: if (isalnum(c) || isblank(c)) {
996: *dst = c;
997: ++dst;
998: ++src;
999: } else {
1000: ++src;
1001: --newlen;
1002: }
1003: }
1004: } else {
1005: wchar_t *dst, *end, *src;
1006: wchar_t c;
1007:
1008: src = str->data.wstr;
1009: dst = src;
1010: end = src + str->len;
1011:
1012: while (src < end) {
1013: c = *src;
1014: if (iswalnum(c) || iswblank(c)) {
1015: *dst = c;
1016: ++dst;
1017: ++src;
1018: } else {
1019: ++src;
1020: --newlen;
1021: }
1022: }
1023: }
1024: bws_setlen(str, newlen);
1025:
1026: return str;
1027: }
1028:
1029: /*
1030: * Converts string to lower case(-f).
1031: */
1032: struct bwstring *
1033: ignore_case(struct bwstring *str)
1034: {
1035: if (sort_mb_cur_max == 1) {
1036: unsigned char *end, *s;
1037:
1038: s = str->data.cstr;
1039: end = s + str->len;
1040:
1041: while (s < end) {
1042: *s = toupper(*s);
1043: ++s;
1044: }
1045: } else {
1046: wchar_t *end, *s;
1047:
1048: s = str->data.wstr;
1049: end = s + str->len;
1050:
1051: while (s < end) {
1052: *s = towupper(*s);
1053: ++s;
1054: }
1055: }
1056: return str;
1057: }
1058:
1059: void
1060: bws_disorder_warnx(struct bwstring *s, const char *fn, size_t pos)
1061: {
1062: if (sort_mb_cur_max == 1)
1063: warnx("%s:%zu: disorder: %s", fn, pos + 1, s->data.cstr);
1064: else
1065: warnx("%s:%zu: disorder: %ls", fn, pos + 1, s->data.wstr);
1066: }