![[BACK]](/icons/back.gif){kind=icon}
Return to coll.c CVS log ![[TXT]](/icons/text.gif){kind=icon}
![[DIR]](/icons/dir.gif){kind=icon}
Up to [local] / src / usr.bin / sort|
Return to coll.c CVS log | Up to [local] / src / usr.bin / sort |
1.7 ! tobias 1: /* $OpenBSD: coll.c,v 1.6 2015/04/01 21:47:19 millert 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 <sys/types.h>
31:
32: #include <errno.h>
33: #include <err.h>
34: #include <langinfo.h>
35: #include <limits.h>
36: #include <math.h>
37: #include <md5.h>
38: #include <stdlib.h>
39: #include <string.h>
40: #include <wchar.h>
41: #include <wctype.h>
42:
43: #include "coll.h"
44: #include "vsort.h"
45:
46: struct key_specs *keys;
47: size_t keys_num = 0;
48:
49: wint_t symbol_decimal_point = L'.';
50: /* there is no default thousands separator in collate rules: */
51: wint_t symbol_thousands_sep = 0;
52: wint_t symbol_negative_sign = L'-';
53: wint_t symbol_positive_sign = L'+';
54:
55: static int wstrcoll(struct key_value *kv1, struct key_value *kv2, size_t offset);
56: static int gnumcoll(struct key_value*, struct key_value *, size_t offset);
57: static int monthcoll(struct key_value*, struct key_value *, size_t offset);
58: static int numcoll(struct key_value*, struct key_value *, size_t offset);
59: static int hnumcoll(struct key_value*, struct key_value *, size_t offset);
60: static int randomcoll(struct key_value*, struct key_value *, size_t offset);
61: static int versioncoll(struct key_value*, struct key_value *, size_t offset);
62:
63: /*
64: * Allocate keys array
65: */
66: struct keys_array *
67: keys_array_alloc(void)
68: {
69: struct keys_array *ka;
70: size_t sz;
71:
72: sz = keys_array_size();
1.2 millert 73: ka = sort_calloc(1, sz);
1.1 millert 74:
75: return ka;
76: }
77:
78: /*
79: * Calculate whether we need key hint space
80: */
81: static size_t
82: key_hint_size(void)
83: {
84: return need_hint ? sizeof(struct key_hint) : 0;
85: }
86:
87: /*
88: * Calculate keys array size
89: */
90: size_t
91: keys_array_size(void)
92: {
93: return keys_num * (sizeof(struct key_value) + key_hint_size());
94: }
95:
96: /*
97: * Clean data of keys array
98: */
99: void
100: clean_keys_array(const struct bwstring *s, struct keys_array *ka)
101: {
102: if (ka) {
103: size_t i;
104:
105: for (i = 0; i < keys_num; ++i)
106: if (ka->key[i].k && ka->key[i].k != s)
107: bwsfree(ka->key[i].k);
108: memset(ka, 0, keys_array_size());
109: }
110: }
111:
112: /*
113: * Set value of a key in the keys set
114: */
115: void
116: set_key_on_keys_array(struct keys_array *ka, struct bwstring *s, size_t ind)
117: {
118: if (ka && keys_num > ind) {
119: struct key_value *kv;
120:
121: kv = &(ka->key[ind]);
122:
1.7 ! tobias 123: if (kv->k != s)
1.1 millert 124: bwsfree(kv->k);
125: kv->k = s;
126: }
127: }
128:
129: /*
130: * Initialize a sort list item
131: */
132: struct sort_list_item *
133: sort_list_item_alloc(void)
134: {
135: struct sort_list_item *si;
136: size_t sz;
137:
138: sz = sizeof(struct sort_list_item) + keys_array_size();
1.2 millert 139: si = sort_calloc(1, sz);
1.1 millert 140:
141: return si;
142: }
143:
144: size_t
145: sort_list_item_size(struct sort_list_item *si)
146: {
147: size_t i, ret = 0;
148:
149: if (si) {
150: ret = sizeof(struct sort_list_item) + keys_array_size();
151: if (si->str)
152: ret += bws_memsize(si->str);
153: for (i = 0; i < keys_num; ++i) {
154: struct key_value *kv;
155:
156: kv = &(si->ka.key[i]);
157:
158: if (kv->k != si->str)
159: ret += bws_memsize(kv->k);
160: }
161: }
162: return ret;
163: }
164:
165: /*
166: * Calculate key for a sort list item
167: */
168: static void
169: sort_list_item_make_key(struct sort_list_item *si)
170: {
171: preproc(si->str, &(si->ka));
172: }
173:
174: /*
175: * Set value of a sort list item.
176: * Return combined string and keys memory size.
177: */
178: void
179: sort_list_item_set(struct sort_list_item *si, struct bwstring *str)
180: {
181: if (si) {
182: clean_keys_array(si->str, &(si->ka));
183: if (si->str) {
184: if (si->str == str) {
185: /* we are trying to reset the same string */
186: return;
187: } else {
188: bwsfree(si->str);
189: si->str = NULL;
190: }
191: }
192: si->str = str;
193: sort_list_item_make_key(si);
194: }
195: }
196:
197: /*
198: * De-allocate a sort list item object memory
199: */
200: void
201: sort_list_item_clean(struct sort_list_item *si)
202: {
203: if (si) {
204: clean_keys_array(si->str, &(si->ka));
205: if (si->str) {
206: bwsfree(si->str);
207: si->str = NULL;
208: }
209: }
210: }
211:
212: /*
213: * Skip columns according to specs
214: */
215: static size_t
216: skip_cols_to_start(const struct bwstring *s, size_t cols, size_t start,
217: bool skip_blanks, bool *empty_key)
218: {
219: if (cols < 1)
220: return BWSLEN(s) + 1;
221:
222: if (skip_blanks)
223: while (start < BWSLEN(s) && iswblank(BWS_GET(s, start)))
224: ++start;
225:
226: while (start < BWSLEN(s) && cols > 1) {
227: --cols;
228: ++start;
229: }
230:
231: if (start >= BWSLEN(s))
232: *empty_key = true;
233:
234: return start;
235: }
236:
237: /*
238: * Skip fields according to specs
239: */
240: static size_t
241: skip_fields_to_start(const struct bwstring *s, size_t fields, bool *empty_field)
242: {
243: if (fields < 2) {
244: if (BWSLEN(s) == 0)
245: *empty_field = true;
246: return 0;
247: } else if (!(sort_opts_vals.tflag)) {
248: size_t cpos = 0;
249: bool pb = true;
250:
251: while (cpos < BWSLEN(s)) {
252: bool isblank;
253:
254: isblank = iswblank(BWS_GET(s, cpos));
255:
256: if (isblank && !pb) {
257: --fields;
258: if (fields <= 1)
259: return cpos;
260: }
261: pb = isblank;
262: ++cpos;
263: }
264: if (fields > 1)
265: *empty_field = true;
266: return cpos;
267: } else {
268: size_t cpos = 0;
269:
270: while (cpos < BWSLEN(s)) {
271: if (BWS_GET(s, cpos) == (wchar_t)sort_opts_vals.field_sep) {
272: --fields;
273: if (fields <= 1)
274: return cpos + 1;
275: }
276: ++cpos;
277: }
278: if (fields > 1)
279: *empty_field = true;
280: return cpos;
281: }
282: }
283:
284: /*
285: * Find fields start
286: */
287: static void
288: find_field_start(const struct bwstring *s, struct key_specs *ks,
289: size_t *field_start, size_t *key_start, bool *empty_field, bool *empty_key)
290: {
291: *field_start = skip_fields_to_start(s, ks->f1, empty_field);
292: if (!*empty_field)
293: *key_start = skip_cols_to_start(s, ks->c1, *field_start,
294: ks->pos1b, empty_key);
295: else
296: *empty_key = true;
297: }
298:
299: /*
300: * Find end key position
301: */
302: static size_t
303: find_field_end(const struct bwstring *s, struct key_specs *ks)
304: {
305: size_t f2, next_field_start, pos_end;
306: bool empty_field, empty_key;
307:
308: empty_field = false;
309: empty_key = false;
310: f2 = ks->f2;
311:
312: if (f2 == 0)
313: return BWSLEN(s) + 1;
314: else {
315: if (ks->c2 == 0) {
316: next_field_start = skip_fields_to_start(s, f2 + 1,
317: &empty_field);
318: if ((next_field_start > 0) && sort_opts_vals.tflag &&
319: ((wchar_t)sort_opts_vals.field_sep == BWS_GET(s,
320: next_field_start - 1)))
321: --next_field_start;
322: } else
323: next_field_start = skip_fields_to_start(s, f2,
324: &empty_field);
325: }
326:
327: if (empty_field || (next_field_start >= BWSLEN(s)))
328: return BWSLEN(s) + 1;
329:
330: if (ks->c2) {
331: pos_end = skip_cols_to_start(s, ks->c2, next_field_start,
332: ks->pos2b, &empty_key);
333: if (pos_end < BWSLEN(s))
334: ++pos_end;
335: } else
336: pos_end = next_field_start;
337:
338: return pos_end;
339: }
340:
341: /*
342: * Cut a field according to the key specs
343: */
344: static struct bwstring *
345: cut_field(const struct bwstring *s, struct key_specs *ks)
346: {
347: struct bwstring *ret = NULL;
348:
349: if (s && ks) {
350: size_t field_start, key_end, key_start, sz;
351: bool empty_field, empty_key;
352:
353: field_start = 0;
354: key_start = 0;
355: empty_field = false;
356: empty_key = false;
357:
358: find_field_start(s, ks, &field_start, &key_start,
359: &empty_field, &empty_key);
360:
361: if (empty_key)
362: sz = 0;
363: else {
364: key_end = find_field_end(s, ks);
365: sz = (key_end < key_start) ? 0 : (key_end - key_start);
366: }
367:
368: ret = bwsalloc(sz);
369: if (sz)
370: bwsnocpy(ret, s, key_start, sz);
371: } else
372: ret = bwsalloc(0);
373:
374: return ret;
375: }
376:
377: /*
378: * Preprocesses a line applying the necessary transformations
379: * specified by command line options and returns the preprocessed
380: * string, which can be used to compare.
381: */
382: int
383: preproc(struct bwstring *s, struct keys_array *ka)
384: {
385: if (sort_opts_vals.kflag) {
386: size_t i;
387: for (i = 0; i < keys_num; i++) {
388: struct bwstring *key;
389: struct key_specs *kspecs;
390: struct sort_mods *sm;
391:
392: kspecs = &(keys[i]);
393: key = cut_field(s, kspecs);
394:
395: sm = &(kspecs->sm);
396: if (sm->dflag)
397: key = dictionary_order(key);
398: else if (sm->iflag)
399: key = ignore_nonprinting(key);
400: if (sm->fflag || sm->Mflag)
401: key = ignore_case(key);
402:
403: set_key_on_keys_array(ka, key, i);
404: }
405: } else {
406: struct bwstring *ret = NULL;
407: struct sort_mods *sm = default_sort_mods;
408:
409: if (sm->bflag) {
410: if (ret == NULL)
411: ret = bwsdup(s);
412: ret = ignore_leading_blanks(ret);
413: }
414: if (sm->dflag) {
415: if (ret == NULL)
416: ret = bwsdup(s);
417: ret = dictionary_order(ret);
418: } else if (sm->iflag) {
419: if (ret == NULL)
420: ret = bwsdup(s);
421: ret = ignore_nonprinting(ret);
422: }
423: if (sm->fflag || sm->Mflag) {
424: if (ret == NULL)
425: ret = bwsdup(s);
426: ret = ignore_case(ret);
427: }
428: if (ret == NULL)
429: set_key_on_keys_array(ka, s, 0);
430: else
431: set_key_on_keys_array(ka, ret, 0);
432: }
433:
434: return 0;
435: }
436:
437: cmpcoll_t
438: get_sort_func(struct sort_mods *sm)
439: {
440: if (sm->nflag)
441: return numcoll;
442: else if (sm->hflag)
443: return hnumcoll;
444: else if (sm->gflag)
445: return gnumcoll;
446: else if (sm->Mflag)
447: return monthcoll;
448: else if (sm->Rflag)
449: return randomcoll;
450: else if (sm->Vflag)
451: return versioncoll;
452: else
453: return wstrcoll;
454: }
455:
456: /*
457: * Compares the given strings. Returns a positive number if
458: * the first precedes the second, a negative number if the second is
459: * the preceding one, and zero if they are equal. This function calls
460: * the underlying collate functions, which done the actual comparison.
461: */
462: int
463: key_coll(struct keys_array *ps1, struct keys_array *ps2, size_t offset)
464: {
465: struct sort_mods *sm;
466: int res = 0;
467: size_t i;
468:
469: for (i = 0; i < keys_num; ++i) {
470: sm = &(keys[i].sm);
471:
472: if (sm->rflag)
473: res = sm->func(&(ps2->key[i]), &(ps1->key[i]), offset);
474: else
475: res = sm->func(&(ps1->key[i]), &(ps2->key[i]), offset);
476:
477: if (res)
478: break;
479:
480: /* offset applies to only the first key */
481: offset = 0;
482: }
483: return res;
484: }
485:
486: /*
487: * Compare two strings.
488: * Plain symbol-by-symbol comparison.
489: */
490: int
491: top_level_str_coll(const struct bwstring *s1, const struct bwstring *s2)
492: {
493: if (default_sort_mods->rflag) {
494: const struct bwstring *tmp;
495:
496: tmp = s1;
497: s1 = s2;
498: s2 = tmp;
499: }
500:
501: return bwscoll(s1, s2, 0);
502: }
503:
504: /*
505: * Compare a string and a sort list item, according to the sort specs.
506: */
507: int
508: str_list_coll(struct bwstring *str1, struct sort_list_item **ss2)
509: {
510: struct keys_array *ka1;
511: int ret = 0;
512:
513: ka1 = keys_array_alloc();
514:
515: preproc(str1, ka1);
516:
517: sort_list_item_make_key(*ss2);
518:
519: if (debug_sort) {
520: bwsprintf(stdout, str1, "; s1=<", ">");
521: bwsprintf(stdout, (*ss2)->str, ", s2=<", ">");
522: }
523:
524: ret = key_coll(ka1, &((*ss2)->ka), 0);
525:
526: if (debug_sort)
527: printf("; cmp1=%d", ret);
528:
529: clean_keys_array(str1, ka1);
530: sort_free(ka1);
531:
532: if ((ret == 0) && !(sort_opts_vals.sflag) && sort_opts_vals.complex_sort) {
533: ret = top_level_str_coll(str1, ((*ss2)->str));
534: if (debug_sort)
535: printf("; cmp2=%d", ret);
536: }
537:
538: if (debug_sort)
539: putchar('\n');
540:
541: return ret;
542: }
543:
544: /*
545: * Compare two sort list items, according to the sort specs.
546: */
547: int
548: list_coll_offset(struct sort_list_item **ss1, struct sort_list_item **ss2,
549: size_t offset)
550: {
551: int ret;
552:
553: ret = key_coll(&((*ss1)->ka), &((*ss2)->ka), offset);
554:
555: if (debug_sort) {
556: if (offset)
557: printf("; offset=%d", (int) offset);
558: bwsprintf(stdout, ((*ss1)->str), "; s1=<", ">");
559: bwsprintf(stdout, ((*ss2)->str), ", s2=<", ">");
560: printf("; cmp1=%d\n", ret);
561: }
562:
563: if (ret)
564: return ret;
565:
566: if (!(sort_opts_vals.sflag) && sort_opts_vals.complex_sort) {
567: ret = top_level_str_coll(((*ss1)->str), ((*ss2)->str));
568: if (debug_sort)
569: printf("; cmp2=%d\n", ret);
570: }
571:
572: return ret;
573: }
574:
575: /*
576: * Compare two sort list items, according to the sort specs.
577: */
578: int
579: list_coll(const void *ss1, const void *ss2)
580: {
581: return list_coll_offset((struct sort_list_item **)ss1,
582: (struct sort_list_item **)ss2, 0);
583: }
584:
585: #define LSCDEF(N) \
586: static int \
587: list_coll_##N(struct sort_list_item **ss1, struct sort_list_item **ss2) \
588: { \
589: \
590: return list_coll_offset(ss1, ss2, N); \
591: }
592:
593: LSCDEF(0)
594: LSCDEF(1)
595: LSCDEF(2)
596: LSCDEF(3)
597: LSCDEF(4)
598: LSCDEF(5)
599: LSCDEF(6)
600: LSCDEF(7)
601: LSCDEF(8)
602: LSCDEF(9)
603: LSCDEF(10)
604: LSCDEF(11)
605: LSCDEF(12)
606: LSCDEF(13)
607: LSCDEF(14)
608: LSCDEF(15)
609: LSCDEF(16)
610: LSCDEF(17)
611: LSCDEF(18)
612: LSCDEF(19)
613: LSCDEF(20)
614:
615: listcoll_t
616: get_list_call_func(size_t offset)
617: {
618: static const listcoll_t lsarray[] = { list_coll_0, list_coll_1,
619: list_coll_2, list_coll_3, list_coll_4, list_coll_5,
620: list_coll_6, list_coll_7, list_coll_8, list_coll_9,
621: list_coll_10, list_coll_11, list_coll_12, list_coll_13,
622: list_coll_14, list_coll_15, list_coll_16, list_coll_17,
623: list_coll_18, list_coll_19, list_coll_20 };
624:
625: if (offset <= 20)
626: return lsarray[offset];
627:
628: return list_coll_0;
629: }
630:
631: /*
632: * Compare two sort list items, only by their original string.
633: */
634: int
635: list_coll_by_str_only(struct sort_list_item **ss1, struct sort_list_item **ss2)
636: {
637: return top_level_str_coll(((*ss1)->str), ((*ss2)->str));
638: }
639:
640: /*
641: * Maximum size of a number in the string (before or after decimal point)
642: */
643: #define MAX_NUM_SIZE (128)
644:
645: /*
646: * Set suffix value
647: */
648: static void
649: setsuffix(wchar_t c, unsigned char *si)
650: {
651: switch (c){
652: case L'k':
653: case L'K':
654: *si = 1;
655: break;
656: case L'M':
657: *si = 2;
658: break;
659: case L'G':
660: *si = 3;
661: break;
662: case L'T':
663: *si = 4;
664: break;
665: case L'P':
666: *si = 5;
667: break;
668: case L'E':
669: *si = 6;
670: break;
671: case L'Z':
672: *si = 7;
673: break;
674: case L'Y':
675: *si = 8;
676: break;
677: default:
678: *si = 0;
679: };
680: }
681:
682: /*
683: * Read string s and parse the string into a fixed-decimal-point number.
684: * sign equals -1 if the number is negative (explicit plus is not allowed,
685: * according to GNU sort's "info sort".
686: * The number part before decimal point is in the smain, after the decimal
687: * point is in sfrac, tail is the pointer to the remainder of the string.
688: */
689: static int
690: read_number(struct bwstring *s0, int *sign, wchar_t *smain, size_t *main_len, wchar_t *sfrac, size_t *frac_len, unsigned char *si)
691: {
692: bwstring_iterator s;
693:
694: s = bws_begin(s0);
695:
696: /* always end the fraction with zero, even if we have no fraction */
697: sfrac[0] = 0;
698:
699: while (iswblank(bws_get_iter_value(s)))
700: s = bws_iterator_inc(s, 1);
701:
702: if (bws_get_iter_value(s) == (wchar_t)symbol_negative_sign) {
703: *sign = -1;
704: s = bws_iterator_inc(s, 1);
705: }
706:
707: // This is '0', not '\0', do not change this
708: while (iswdigit(bws_get_iter_value(s)) &&
709: (bws_get_iter_value(s) == L'0'))
710: s = bws_iterator_inc(s, 1);
711:
712: while (bws_get_iter_value(s) && *main_len < MAX_NUM_SIZE) {
713: if (iswdigit(bws_get_iter_value(s))) {
714: smain[*main_len] = bws_get_iter_value(s);
715: s = bws_iterator_inc(s, 1);
716: *main_len += 1;
717: } else if (symbol_thousands_sep &&
718: (bws_get_iter_value(s) == (wchar_t)symbol_thousands_sep))
719: s = bws_iterator_inc(s, 1);
720: else
721: break;
722: }
723:
724: smain[*main_len] = 0;
725:
726: if (bws_get_iter_value(s) == (wchar_t)symbol_decimal_point) {
727: s = bws_iterator_inc(s, 1);
728: while (iswdigit(bws_get_iter_value(s)) &&
729: *frac_len < MAX_NUM_SIZE) {
730: sfrac[*frac_len] = bws_get_iter_value(s);
731: s = bws_iterator_inc(s, 1);
732: *frac_len += 1;
733: }
734: sfrac[*frac_len] = 0;
735:
736: while (*frac_len > 0 && sfrac[*frac_len - 1] == L'0') {
737: --(*frac_len);
738: sfrac[*frac_len] = L'\0';
739: }
740: }
741:
742: setsuffix(bws_get_iter_value(s), si);
743:
744: if ((*main_len + *frac_len) == 0)
745: *sign = 0;
746:
747: return 0;
748: }
749:
750: /*
751: * Implements string sort.
752: */
753: static int
754: wstrcoll(struct key_value *kv1, struct key_value *kv2, size_t offset)
755: {
756:
757: if (debug_sort) {
758: if (offset)
759: printf("; offset=%d\n", (int) offset);
760: bwsprintf(stdout, kv1->k, "; k1=<", ">");
761: printf("(%zu)", BWSLEN(kv1->k));
762: bwsprintf(stdout, kv2->k, ", k2=<", ">");
763: printf("(%zu)", BWSLEN(kv2->k));
764: }
765:
766: return bwscoll(kv1->k, kv2->k, offset);
767: }
768:
769: /*
770: * Compare two suffixes
771: */
772: static inline int
773: cmpsuffix(unsigned char si1, unsigned char si2)
774: {
775: return (char)si1 - (char)si2;
776: }
777:
778: /*
779: * Implements numeric sort for -n and -h.
780: */
781: static int
782: numcoll_impl(struct key_value *kv1, struct key_value *kv2,
783: size_t offset __unused, bool use_suffix)
784: {
785: struct bwstring *s1, *s2;
786: wchar_t sfrac1[MAX_NUM_SIZE + 1], sfrac2[MAX_NUM_SIZE + 1];
787: wchar_t smain1[MAX_NUM_SIZE + 1], smain2[MAX_NUM_SIZE + 1];
788: int cmp_res, sign1, sign2;
789: size_t frac1, frac2, main1, main2;
790: unsigned char SI1, SI2;
791: bool e1, e2, key1_read, key2_read;
792:
793: s1 = kv1->k;
794: s2 = kv2->k;
795: sign1 = sign2 = 0;
796: main1 = main2 = 0;
797: frac1 = frac2 = 0;
798:
799: key1_read = key2_read = false;
800:
801: if (debug_sort) {
802: bwsprintf(stdout, s1, "; k1=<", ">");
803: bwsprintf(stdout, s2, ", k2=<", ">");
804: }
805:
806: if (s1 == s2)
807: return 0;
808:
809: if (kv1->hint->status == HS_UNINITIALIZED) {
810: /* read the number from the string */
811: read_number(s1, &sign1, smain1, &main1, sfrac1, &frac1, &SI1);
812: key1_read = true;
813: kv1->hint->v.nh.n1 = wcstoull(smain1, NULL, 10);
814: if (main1 < 1 && frac1 < 1)
815: kv1->hint->v.nh.empty=true;
816: kv1->hint->v.nh.si = SI1;
817: kv1->hint->status = (kv1->hint->v.nh.n1 != ULLONG_MAX) ?
818: HS_INITIALIZED : HS_ERROR;
819: kv1->hint->v.nh.neg = (sign1 < 0) ? true : false;
820: }
821:
822: if (kv2->hint->status == HS_UNINITIALIZED) {
823: /* read the number from the string */
824: read_number(s2, &sign2, smain2, &main2, sfrac2, &frac2, &SI2);
825: key2_read = true;
826: kv2->hint->v.nh.n1 = wcstoull(smain2, NULL, 10);
827: if (main2 < 1 && frac2 < 1)
828: kv2->hint->v.nh.empty=true;
829: kv2->hint->v.nh.si = SI2;
830: kv2->hint->status = (kv2->hint->v.nh.n1 != ULLONG_MAX) ?
831: HS_INITIALIZED : HS_ERROR;
832: kv2->hint->v.nh.neg = (sign2 < 0) ? true : false;
833: }
834:
835: if (kv1->hint->status == HS_INITIALIZED && kv2->hint->status ==
836: HS_INITIALIZED) {
837: unsigned long long n1, n2;
838: bool neg1, neg2;
839:
840: e1 = kv1->hint->v.nh.empty;
841: e2 = kv2->hint->v.nh.empty;
842:
843: if (e1 && e2)
844: return 0;
845:
846: neg1 = kv1->hint->v.nh.neg;
847: neg2 = kv2->hint->v.nh.neg;
848:
849: if (neg1 && !neg2)
850: return -1;
851: if (neg2 && !neg1)
852: return 1;
853:
854: if (e1)
855: return neg2 ? 1 : -1;
856: else if (e2)
857: return neg1 ? -1 : 1;
858:
859:
860: if (use_suffix) {
861: cmp_res = cmpsuffix(kv1->hint->v.nh.si, kv2->hint->v.nh.si);
862: if (cmp_res)
863: return neg1 ? -cmp_res : cmp_res;
864: }
865:
866: n1 = kv1->hint->v.nh.n1;
867: n2 = kv2->hint->v.nh.n1;
868: if (n1 < n2)
869: return neg1 ? 1 : -1;
870: else if (n1 > n2)
871: return neg1 ? -1 : 1;
872: }
873:
874: /* read the numbers from the strings */
875: if (!key1_read)
876: read_number(s1, &sign1, smain1, &main1, sfrac1, &frac1, &SI1);
877: if (!key2_read)
878: read_number(s2, &sign2, smain2, &main2, sfrac2, &frac2, &SI2);
879:
880: e1 = ((main1 + frac1) == 0);
881: e2 = ((main2 + frac2) == 0);
882:
883: if (e1 && e2)
884: return 0;
885:
886: /* we know the result if the signs are different */
887: if (sign1 < 0 && sign2 >= 0)
888: return -1;
889: if (sign1 >= 0 && sign2 < 0)
890: return 1;
891:
892: if (e1)
893: return (sign2 < 0) ? +1 : -1;
894: else if (e2)
895: return (sign1 < 0) ? -1 : 1;
896:
897: if (use_suffix) {
898: cmp_res = cmpsuffix(SI1, SI2);
899: if (cmp_res)
900: return (sign1 < 0) ? -cmp_res : cmp_res;
901: }
902:
903: /* if both numbers are empty assume that the strings are equal */
904: if (main1 < 1 && main2 < 1 && frac1 < 1 && frac2 < 1)
905: return 0;
906:
907: /*
908: * if the main part is of different size, we know the result
909: * (because the leading zeros are removed)
910: */
911: if (main1 < main2)
912: cmp_res = -1;
913: else if (main1 > main2)
914: cmp_res = +1;
915: /* if the sizes are equal then simple non-collate string compare gives the correct result */
916: else
917: cmp_res = wcscmp(smain1, smain2);
918:
919: /* check fraction */
920: if (!cmp_res)
921: cmp_res = wcscmp(sfrac1, sfrac2);
922:
923: if (!cmp_res)
924: return 0;
925:
926: /* reverse result if the signs are negative */
927: if (sign1 < 0 && sign2 < 0)
928: cmp_res = -cmp_res;
929:
930: return cmp_res;
931: }
932:
933: /*
934: * Implements numeric sort (-n).
935: */
936: static int
937: numcoll(struct key_value *kv1, struct key_value *kv2, size_t offset)
938: {
939: return numcoll_impl(kv1, kv2, offset, false);
940: }
941:
942: /*
943: * Implements 'human' numeric sort (-h).
944: */
945: static int
946: hnumcoll(struct key_value *kv1, struct key_value *kv2, size_t offset)
947: {
948: return numcoll_impl(kv1, kv2, offset, true);
949: }
950:
951: /*
952: * Implements random sort (-R).
953: */
954: static int
955: randomcoll(struct key_value *kv1, struct key_value *kv2,
956: size_t offset __unused)
957: {
958: struct bwstring *s1, *s2;
959: MD5_CTX ctx1, ctx2;
960: char *b1, *b2;
961:
962: s1 = kv1->k;
963: s2 = kv2->k;
964:
965: if (debug_sort) {
966: bwsprintf(stdout, s1, "; k1=<", ">");
967: bwsprintf(stdout, s2, ", k2=<", ">");
968: }
969:
970: if (s1 == s2)
971: return 0;
972:
973: memcpy(&ctx1, &md5_ctx, sizeof(MD5_CTX));
974: memcpy(&ctx2, &md5_ctx, sizeof(MD5_CTX));
975:
976: MD5Update(&ctx1, bwsrawdata(s1), bwsrawlen(s1));
977: MD5Update(&ctx2, bwsrawdata(s2), bwsrawlen(s2));
978: b1 = MD5End(&ctx1, NULL);
979: b2 = MD5End(&ctx2, NULL);
980: if (b1 == NULL) {
981: if (b2 == NULL)
982: return 0;
983: else {
984: sort_free(b2);
985: return -1;
986: }
987: } else if (b2 == NULL) {
988: sort_free(b1);
989: return 1;
990: } else {
991: int cmp_res;
992:
993: cmp_res = strcmp(b1, b2);
994: sort_free(b1);
995: sort_free(b2);
996:
997: if (!cmp_res)
998: cmp_res = bwscoll(s1, s2, 0);
999:
1000: return cmp_res;
1001: }
1002: }
1003:
1004: /*
1005: * Implements version sort (-V).
1006: */
1007: static int
1008: versioncoll(struct key_value *kv1, struct key_value *kv2,
1009: size_t offset __unused)
1010: {
1011: struct bwstring *s1, *s2;
1012:
1013: s1 = kv1->k;
1014: s2 = kv2->k;
1015:
1016: if (debug_sort) {
1017: bwsprintf(stdout, s1, "; k1=<", ">");
1018: bwsprintf(stdout, s2, ", k2=<", ">");
1019: }
1020:
1021: if (s1 == s2)
1022: return 0;
1023:
1024: return vcmp(s1, s2);
1025: }
1026:
1027: /*
1028: * Check for minus infinity
1029: */
1030: static inline bool
1031: huge_minus(double d, int err1)
1032: {
1033: if (err1 == ERANGE)
1034: if (d == -HUGE_VAL || d == -HUGE_VALF || d == -HUGE_VALL)
1035: return 1;
1036:
1037: return 0;
1038: }
1039:
1040: /*
1041: * Check for plus infinity
1042: */
1043: static inline bool
1044: huge_plus(double d, int err1)
1045: {
1046: if (err1 == ERANGE)
1047: if (d == HUGE_VAL || d == HUGE_VALF || d == HUGE_VALL)
1048: return 1;
1049:
1050: return 0;
1051: }
1052:
1053: /*
1054: * Check whether a function is a NAN
1055: */
1056: static bool
1057: is_nan(double d)
1058: {
1.3 miod 1059: #if defined(NAN)
1.1 millert 1060: return (d == NAN || isnan(d));
1.3 miod 1061: #else
1062: return (isnan(d));
1063: #endif
1.1 millert 1064: }
1065:
1066: /*
1067: * Compare two NANs
1068: */
1069: static int
1070: cmp_nans(double d1, double d2)
1071: {
1072: if (d1 == d2)
1073: return 0;
1074: return d1 < d2 ? -1 : 1;
1075: }
1076:
1077: /*
1078: * Implements general numeric sort (-g).
1079: */
1080: static int
1081: gnumcoll(struct key_value *kv1, struct key_value *kv2,
1082: size_t offset __unused)
1083: {
1084: double d1, d2;
1085: int err1, err2;
1086: bool empty1, empty2, key1_read, key2_read;
1087:
1088: d1 = d2 = 0;
1089: err1 = err2 = 0;
1090: key1_read = key2_read = false;
1091:
1092: if (debug_sort) {
1093: bwsprintf(stdout, kv1->k, "; k1=<", ">");
1094: bwsprintf(stdout, kv2->k, "; k2=<", ">");
1095: }
1096:
1097: if (kv1->hint->status == HS_UNINITIALIZED) {
1098: errno = 0;
1099: d1 = bwstod(kv1->k, &empty1);
1100: err1 = errno;
1101:
1102: if (empty1)
1103: kv1->hint->v.gh.notnum = true;
1104: else if (err1 == 0) {
1105: kv1->hint->v.gh.d = d1;
1106: kv1->hint->v.gh.nan = is_nan(d1);
1107: kv1->hint->status = HS_INITIALIZED;
1108: } else
1109: kv1->hint->status = HS_ERROR;
1110:
1111: key1_read = true;
1112: }
1113:
1114: if (kv2->hint->status == HS_UNINITIALIZED) {
1115: errno = 0;
1116: d2 = bwstod(kv2->k, &empty2);
1117: err2 = errno;
1118:
1119: if (empty2)
1120: kv2->hint->v.gh.notnum = true;
1121: else if (err2 == 0) {
1122: kv2->hint->v.gh.d = d2;
1123: kv2->hint->v.gh.nan = is_nan(d2);
1124: kv2->hint->status = HS_INITIALIZED;
1125: } else
1126: kv2->hint->status = HS_ERROR;
1127:
1128: key2_read = true;
1129: }
1130:
1131: if (kv1->hint->status == HS_INITIALIZED &&
1132: kv2->hint->status == HS_INITIALIZED) {
1133: if (kv1->hint->v.gh.notnum)
1134: return kv2->hint->v.gh.notnum ? 0 : -1;
1135: else if (kv2->hint->v.gh.notnum)
1136: return 1;
1137:
1138: if (kv1->hint->v.gh.nan)
1139: return kv2->hint->v.gh.nan ?
1140: cmp_nans(kv1->hint->v.gh.d, kv2->hint->v.gh.d) : -1;
1141: else if (kv2->hint->v.gh.nan)
1142: return 1;
1143:
1144: d1 = kv1->hint->v.gh.d;
1145: d2 = kv2->hint->v.gh.d;
1146:
1147: if (d1 < d2)
1148: return -1;
1149: else if (d1 > d2)
1150: return 1;
1151: else
1152: return 0;
1153: }
1154:
1155: if (!key1_read) {
1156: errno = 0;
1157: d1 = bwstod(kv1->k, &empty1);
1158: err1 = errno;
1159: }
1160:
1161: if (!key2_read) {
1162: errno = 0;
1163: d2 = bwstod(kv2->k, &empty2);
1164: err2 = errno;
1165: }
1166:
1167: /* Non-value case: */
1168: if (empty1)
1169: return empty2 ? 0 : -1;
1170: else if (empty2)
1171: return 1;
1172:
1173: /* NAN case */
1174: if (is_nan(d1))
1175: return is_nan(d2) ? cmp_nans(d1, d2) : -1;
1176: else if (is_nan(d2))
1177: return 1;
1178:
1179: /* Infinities */
1180: if (err1 == ERANGE || err2 == ERANGE) {
1181: /* Minus infinity case */
1182: if (huge_minus(d1, err1)) {
1183: if (huge_minus(d2, err2)) {
1184: if (d1 == d2)
1185: return 0;
1186: return d1 < d2 ? -1 : 1;
1187: } else
1188: return -1;
1189:
1190: } else if (huge_minus(d2, err2)) {
1191: if (huge_minus(d1, err1)) {
1192: if (d1 == d2)
1193: return 0;
1194: return d1 < d2 ? -1 : 1;
1195: } else
1196: return 1;
1197: }
1198:
1199: /* Plus infinity case */
1200: if (huge_plus(d1, err1)) {
1201: if (huge_plus(d2, err2)) {
1202: if (d1 == d2)
1203: return 0;
1204: return d1 < d2 ? -1 : 1;
1205: } else
1206: return 1;
1207: } else if (huge_plus(d2, err2)) {
1208: if (huge_plus(d1, err1)) {
1209: if (d1 == d2)
1210: return 0;
1211: return d1 < d2 ? -1 : 1;
1212: } else
1213: return -1;
1214: }
1215: }
1216:
1217: if (d1 == d2)
1218: return 0;
1219: return d1 < d2 ? -1 : 1;
1220: }
1221:
1222: /*
1223: * Implements month sort (-M).
1224: */
1225: static int
1226: monthcoll(struct key_value *kv1, struct key_value *kv2, size_t offset __unused)
1227: {
1228: int val1, val2;
1229: bool key1_read, key2_read;
1230:
1231: val1 = val2 = 0;
1232: key1_read = key2_read = false;
1233:
1234: if (debug_sort) {
1235: bwsprintf(stdout, kv1->k, "; k1=<", ">");
1236: bwsprintf(stdout, kv2->k, "; k2=<", ">");
1237: }
1238:
1239: if (kv1->hint->status == HS_UNINITIALIZED) {
1240: kv1->hint->v.Mh.m = bws_month_score(kv1->k);
1241: key1_read = true;
1242: kv1->hint->status = HS_INITIALIZED;
1243: }
1244:
1245: if (kv2->hint->status == HS_UNINITIALIZED) {
1246: kv2->hint->v.Mh.m = bws_month_score(kv2->k);
1247: key2_read = true;
1248: kv2->hint->status = HS_INITIALIZED;
1249: }
1250:
1251: if (kv1->hint->status == HS_INITIALIZED) {
1252: val1 = kv1->hint->v.Mh.m;
1253: key1_read = true;
1254: }
1255:
1256: if (kv2->hint->status == HS_INITIALIZED) {
1257: val2 = kv2->hint->v.Mh.m;
1258: key2_read = true;
1259: }
1260:
1261: if (!key1_read)
1262: val1 = bws_month_score(kv1->k);
1263: if (!key2_read)
1264: val2 = bws_month_score(kv2->k);
1265:
1266: if (val1 == val2)
1267: return 0;
1268: return val1 < val2 ? -1 : 1;
1269: }