aboutsummaryrefslogtreecommitdiff
path: root/coreutils/diff.c
blob: 072e4bc208e19f6e7a2839aa3f6b486e0488a0b2 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
/* vi: set sw=4 ts=4: */
/*
 * Mini diff implementation for busybox, adapted from OpenBSD diff.
 *
 * Copyright (C) 2006 by Robert Sullivan <cogito.ergo.cogito@hotmail.com>
 * Copyright (c) 2003 Todd C. Miller <Todd.Miller@courtesan.com>
 *
 * Sponsored in part by the Defense Advanced Research Projects
 * Agency (DARPA) and Air Force Research Laboratory, Air Force
 * Materiel Command, USAF, under agreement number F39502-99-1-0512.
 *
 * Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
 */

#include "busybox.h"

#define FSIZE_MAX 32768

/*
 * Output flags
 */
#define D_HEADER        1	/* Print a header/footer between files */
#define D_EMPTY1        2	/* Treat first file as empty (/dev/null) */
#define D_EMPTY2        4	/* Treat second file as empty (/dev/null) */

/*
 * Status values for print_status() and diffreg() return values
 * Guide:
 * D_SAME - files are the same
 * D_DIFFER - files differ
 * D_BINARY - binary files differ
 * D_COMMON - subdirectory common to both dirs
 * D_ONLY - file only exists in one dir
 * D_MISMATCH1 - path1 a dir, path2 a file
 * D_MISMATCH2 - path1 a file, path2 a dir
 * D_ERROR - error occurred
 * D_SKIPPED1 - skipped path1 as it is a special file
 * D_SKIPPED2 - skipped path2 as it is a special file
 */

#define D_SAME		0
#define D_DIFFER	(1<<0)
#define D_BINARY	(1<<1)
#define D_COMMON	(1<<2)
#define D_ONLY		(1<<3)
#define D_MISMATCH1	(1<<4)
#define D_MISMATCH2	(1<<5)
#define D_ERROR		(1<<6)
#define D_SKIPPED1	(1<<7)
#define D_SKIPPED2	(1<<8)

/* Command line options */
#define FLAG_a	(1<<0)
#define FLAG_b	(1<<1)
#define FLAG_d  (1<<2)
#define FLAG_i	(1<<3)
#define FLAG_L	(1<<4)
#define FLAG_N	(1<<5)
#define FLAG_q	(1<<6)
#define FLAG_r	(1<<7)
#define FLAG_s	(1<<8)
#define FLAG_S	(1<<9)
#define FLAG_t	(1<<10)
#define FLAG_T	(1<<11)
#define FLAG_U	(1<<12)
#define	FLAG_w	(1<<13)

/* XXX: FIXME: the following variables should be static, but gcc currently
 * creates a much bigger object if we do this. [which version of gcc? --vda] */
int context, status;
char *start, *label[2];
struct stat stb1, stb2;
char **dl;
USE_FEATURE_DIFF_DIR(static int dl_count;)

struct cand {
	int x;
	int y;
	int pred;
};

struct line {
	int serial;
	int value;
} *file[2];

/*
 * The following struct is used to record change information
 * doing a "context" or "unified" diff.  (see routine "change" to
 * understand the highly mnemonic field names)
 */
struct context_vec {
	int a;				/* start line in old file */
	int b;				/* end line in old file */
	int c;				/* start line in new file */
	int d;				/* end line in new file */
};

static int *J;			/* will be overlaid on class */
static int *class;		/* will be overlaid on file[0] */
static int *klist;		/* will be overlaid on file[0] after class */
static int *member;		/* will be overlaid on file[1] */
static int clen;
static int len[2];
static int pref, suff;	/* length of prefix and suffix */
static int slen[2];
static int anychange;
static long *ixnew;		/* will be overlaid on file[1] */
static long *ixold;		/* will be overlaid on klist */
static struct cand *clist;	/* merely a free storage pot for candidates */
static int clistlen;	/* the length of clist */
static struct line *sfile[2];	/* shortened by pruning common prefix/suffix */
static struct context_vec *context_vec_start;
static struct context_vec *context_vec_end;
static struct context_vec *context_vec_ptr;


static void print_only(const char *path, size_t dirlen, const char *entry)
{
	if (dirlen > 1)
		dirlen--;
	printf("Only in %.*s: %s\n", (int) dirlen, path, entry);
}


static void print_status(int val, char *path1, char *path2, char *entry)
{
	const char *const _entry = entry ? entry : "";
	char *_path1 = entry ? concat_path_file(path1, _entry) : path1;
	char *_path2 = entry ? concat_path_file(path2, _entry) : path2;

	switch (val) {
	case D_ONLY:
		print_only(path1, strlen(path1), entry);
		break;
	case D_COMMON:
		printf("Common subdirectories: %s and %s\n", _path1, _path2);
		break;
	case D_BINARY:
		printf("Binary files %s and %s differ\n", _path1, _path2);
		break;
	case D_DIFFER:
		if (option_mask32 & FLAG_q)
			printf("Files %s and %s differ\n", _path1, _path2);
		break;
	case D_SAME:
		if (option_mask32 & FLAG_s)
			printf("Files %s and %s are identical\n", _path1, _path2);
		break;
	case D_MISMATCH1:
		printf("File %s is a %s while file %s is a %s\n",
			   _path1, "directory", _path2, "regular file");
		break;
	case D_MISMATCH2:
		printf("File %s is a %s while file %s is a %s\n",
			   _path1, "regular file", _path2, "directory");
		break;
	case D_SKIPPED1:
		printf("File %s is not a regular file or directory and was skipped\n",
			   _path1);
		break;
	case D_SKIPPED2:
		printf("File %s is not a regular file or directory and was skipped\n",
			   _path2);
		break;
	}
	if (entry) {
		free(_path1);
		free(_path2);
	}
}


/*
 * Hash function taken from Robert Sedgewick, Algorithms in C, 3d ed., p 578.
 */
static int readhash(FILE * f)
{
	int i, t, space;
	int sum;

	sum = 1;
	space = 0;
	if (!(option_mask32 & FLAG_b) && !(option_mask32 & FLAG_w)) {
		if (FLAG_i)
			for (i = 0; (t = getc(f)) != '\n'; i++) {
				if (t == EOF) {
					if (i == 0)
						return 0;
					break;
				}
				sum = sum * 127 + t;
		} else
			for (i = 0; (t = getc(f)) != '\n'; i++) {
				if (t == EOF) {
					if (i == 0)
						return 0;
					break;
				}
				sum = sum * 127 + t;
			}
	} else {
		for (i = 0;;) {
			switch (t = getc(f)) {
			case '\t':
			case '\r':
			case '\v':
			case '\f':
			case ' ':
				space++;
				continue;
			default:
				if (space && !(option_mask32 & FLAG_w)) {
					i++;
					space = 0;
				}
				sum = sum * 127 + t;
				i++;
				continue;
			case EOF:
				if (i == 0)
					return 0;
				/* FALLTHROUGH */
			case '\n':
				break;
			}
			break;
		}
	}
	/*
	 * There is a remote possibility that we end up with a zero sum.
	 * Zero is used as an EOF marker, so return 1 instead.
	 */
	return (sum == 0 ? 1 : sum);
}


/*
 * Check to see if the given files differ.
 * Returns 0 if they are the same, 1 if different, and -1 on error.
 */
static int files_differ(FILE * f1, FILE * f2, int flags)
{
	size_t i, j;

	if ((flags & (D_EMPTY1 | D_EMPTY2)) || stb1.st_size != stb2.st_size ||
		(stb1.st_mode & S_IFMT) != (stb2.st_mode & S_IFMT))
		return 1;
	while (1) {
		i = fread(bb_common_bufsiz1,            1, BUFSIZ/2, f1);
		j = fread(bb_common_bufsiz1 + BUFSIZ/2, 1, BUFSIZ/2, f2);
		if (i != j)
			return 1;
		if (i == 0 && j == 0) {
			if (ferror(f1) || ferror(f2))
				return 1;
			return 0;
		}
		if (memcmp(bb_common_bufsiz1,
		           bb_common_bufsiz1 + BUFSIZ/2, i) != 0)
			return 1;
	}
}


static void prepare(int i, FILE * fd, off_t filesize)
{
	struct line *p;
	int h;
	size_t j, sz;

	rewind(fd);

	sz = (filesize <= FSIZE_MAX ? filesize : FSIZE_MAX) / 25;
	if (sz < 100)
		sz = 100;

	p = xmalloc((sz + 3) * sizeof(struct line));
	for (j = 0; (h = readhash(fd));) {
		if (j == sz) {
			sz = sz * 3 / 2;
			p = xrealloc(p, (sz + 3) * sizeof(struct line));
		}
		p[++j].value = h;
	}
	len[i] = j;
	file[i] = p;
}


static void prune(void)
{
	int i, j;

	for (pref = 0; pref < len[0] && pref < len[1] &&
		 file[0][pref + 1].value == file[1][pref + 1].value; pref++);
	for (suff = 0; suff < len[0] - pref && suff < len[1] - pref &&
		 file[0][len[0] - suff].value == file[1][len[1] - suff].value;
		 suff++);
	for (j = 0; j < 2; j++) {
		sfile[j] = file[j] + pref;
		slen[j] = len[j] - pref - suff;
		for (i = 0; i <= slen[j]; i++)
			sfile[j][i].serial = i;
	}
}


static void equiv(struct line *a, int n, struct line *b, int m, int *c)
{
	int i, j;

	i = j = 1;
	while (i <= n && j <= m) {
		if (a[i].value < b[j].value)
			a[i++].value = 0;
		else if (a[i].value == b[j].value)
			a[i++].value = j;
		else
			j++;
	}
	while (i <= n)
		a[i++].value = 0;
	b[m + 1].value = 0;
	j = 0;
	while (++j <= m) {
		c[j] = -b[j].serial;
		while (b[j + 1].value == b[j].value) {
			j++;
			c[j] = b[j].serial;
		}
	}
	c[j] = -1;
}


static int isqrt(int n)
{
	int y, x = 1;

	if (n == 0)
		return 0;

	do {
		y = x;
		x = n / x;
		x += y;
		x /= 2;
	} while ((x - y) > 1 || (x - y) < -1);

	return x;
}


static int newcand(int x, int y, int pred)
{
	struct cand *q;

	if (clen == clistlen) {
		clistlen = clistlen * 11 / 10;
		clist = xrealloc(clist, clistlen * sizeof(struct cand));
	}
	q = clist + clen;
	q->x = x;
	q->y = y;
	q->pred = pred;
	return clen++;
}


static int search(int *c, int k, int y)
{
	int i, j, l, t;

	if (clist[c[k]].y < y)	/* quick look for typical case */
		return k + 1;
	i = 0;
	j = k + 1;
	while (1) {
		l = i + j;
		if ((l >>= 1) <= i)
			break;
		t = clist[c[l]].y;
		if (t > y)
			j = l;
		else if (t < y)
			i = l;
		else
			return l;
	}
	return l + 1;
}


static int stone(int *a, int n, int *b, int *c)
{
	int i, k, y, j, l;
	int oldc, tc, oldl;
	unsigned int numtries;

#if ENABLE_FEATURE_DIFF_MINIMAL
	const unsigned int bound =
		(option_mask32 & FLAG_d) ? UINT_MAX : MAX(256, isqrt(n));
#else
	const unsigned int bound = MAX(256, isqrt(n));
#endif
	k = 0;
	c[0] = newcand(0, 0, 0);
	for (i = 1; i <= n; i++) {
		j = a[i];
		if (j == 0)
			continue;
		y = -b[j];
		oldl = 0;
		oldc = c[0];
		numtries = 0;
		do {
			if (y <= clist[oldc].y)
				continue;
			l = search(c, k, y);
			if (l != oldl + 1)
				oldc = c[l - 1];
			if (l <= k) {
				if (clist[c[l]].y <= y)
					continue;
				tc = c[l];
				c[l] = newcand(i, y, oldc);
				oldc = tc;
				oldl = l;
				numtries++;
			} else {
				c[l] = newcand(i, y, oldc);
				k++;
				break;
			}
		} while ((y = b[++j]) > 0 && numtries < bound);
	}
	return k;
}


static void unravel(int p)
{
	struct cand *q;
	int i;

	for (i = 0; i <= len[0]; i++)
		J[i] = i <= pref ? i : i > len[0] - suff ? i + len[1] - len[0] : 0;
	for (q = clist + p; q->y != 0; q = clist + q->pred)
		J[q->x + pref] = q->y + pref;
}


static void unsort(struct line *f, int l, int *b)
{
	int *a, i;

	a = xmalloc((l + 1) * sizeof(int));
	for (i = 1; i <= l; i++)
		a[f[i].serial] = f[i].value;
	for (i = 1; i <= l; i++)
		b[i] = a[i];
	free(a);
}


static int skipline(FILE * f)
{
	int i, c;

	for (i = 1; (c = getc(f)) != '\n' && c != EOF; i++)
		continue;
	return i;
}


/*
 * Check does double duty:
 *  1.  ferret out any fortuitous correspondences due
 *      to confounding by hashing (which result in "jackpot")
 *  2.  collect random access indexes to the two files
 */
static void check(FILE * f1, FILE * f2)
{
	int i, j, jackpot, c, d;
	long ctold, ctnew;

	rewind(f1);
	rewind(f2);
	j = 1;
	ixold[0] = ixnew[0] = 0;
	jackpot = 0;
	ctold = ctnew = 0;
	for (i = 1; i <= len[0]; i++) {
		if (J[i] == 0) {
			ixold[i] = ctold += skipline(f1);
			continue;
		}
		while (j < J[i]) {
			ixnew[j] = ctnew += skipline(f2);
			j++;
		}
		if ((option_mask32 & FLAG_b) || (option_mask32 & FLAG_w)
			|| (option_mask32 & FLAG_i)) {
			while (1) {
				c = getc(f1);
				d = getc(f2);
				/*
				 * GNU diff ignores a missing newline
				 * in one file if bflag || wflag.
				 */
				if (((option_mask32 & FLAG_b) || (option_mask32 & FLAG_w)) &&
					((c == EOF && d == '\n') || (c == '\n' && d == EOF))) {
					break;
				}
				ctold++;
				ctnew++;
				if ((option_mask32 & FLAG_b) && isspace(c) && isspace(d)) {
					do {
						if (c == '\n')
							break;
						ctold++;
					} while (isspace(c = getc(f1)));
					do {
						if (d == '\n')
							break;
						ctnew++;
					} while (isspace(d = getc(f2)));
				} else if (option_mask32 & FLAG_w) {
					while (isspace(c) && c != '\n') {
						c = getc(f1);
						ctold++;
					}
					while (isspace(d) && d != '\n') {
						d = getc(f2);
						ctnew++;
					}
				}
				if (c != d) {
					jackpot++;
					J[i] = 0;
					if (c != '\n' && c != EOF)
						ctold += skipline(f1);
					if (d != '\n' && c != EOF)
						ctnew += skipline(f2);
					break;
				}
				if (c == '\n' || c == EOF)
					break;
			}
		} else {
			while (1) {
				ctold++;
				ctnew++;
				if ((c = getc(f1)) != (d = getc(f2))) {
					J[i] = 0;
					if (c != '\n' && c != EOF)
						ctold += skipline(f1);
					if (d != '\n' && c != EOF)
						ctnew += skipline(f2);
					break;
				}
				if (c == '\n' || c == EOF)
					break;
			}
		}
		ixold[i] = ctold;
		ixnew[j] = ctnew;
		j++;
	}
	for (; j <= len[1]; j++)
		ixnew[j] = ctnew += skipline(f2);
}


/* shellsort CACM #201 */
static void sort(struct line *a, int n)
{
	struct line *ai, *aim, w;
	int j, m = 0, k;

	if (n == 0)
		return;
	for (j = 1; j <= n; j *= 2)
		m = 2 * j - 1;
	for (m /= 2; m != 0; m /= 2) {
		k = n - m;
		for (j = 1; j <= k; j++) {
			for (ai = &a[j]; ai > a; ai -= m) {
				aim = &ai[m];
				if (aim < ai)
					break;	/* wraparound */
				if (aim->value > ai[0].value ||
					(aim->value == ai[0].value && aim->serial > ai[0].serial))
					break;
				w.value = ai[0].value;
				ai[0].value = aim->value;
				aim->value = w.value;
				w.serial = ai[0].serial;
				ai[0].serial = aim->serial;
				aim->serial = w.serial;
			}
		}
	}
}


static void uni_range(int a, int b)
{
	if (a < b)
		printf("%d,%d", a, b - a + 1);
	else if (a == b)
		printf("%d", b);
	else
		printf("%d,0", b);
}


static int fetch(long *f, int a, int b, FILE * lb, int ch)
{
	int i, j, c, lastc, col, nc;

	if (a > b)
		return 0;
	for (i = a; i <= b; i++) {
		fseek(lb, f[i - 1], SEEK_SET);
		nc = f[i] - f[i - 1];
		if (ch != '\0') {
			putchar(ch);
			if (option_mask32 & FLAG_T)
				putchar('\t');
		}
		col = 0;
		for (j = 0, lastc = '\0'; j < nc; j++, lastc = c) {
			if ((c = getc(lb)) == EOF) {
				puts("\n\\ No newline at end of file");
				return 0;
			}
			if (c == '\t' && (option_mask32 & FLAG_t)) {
				do {
					putchar(' ');
				} while (++col & 7);
			} else {
				putchar(c);
				col++;
			}
		}
	}
	return 0;
}


static int asciifile(FILE * f)
{
#if ENABLE_FEATURE_DIFF_BINARY
	int i, cnt;
#endif

	if ((option_mask32 & FLAG_a) || f == NULL)
		return 1;

#if ENABLE_FEATURE_DIFF_BINARY
	rewind(f);
	cnt = fread(bb_common_bufsiz1, 1, BUFSIZ, f);
	for (i = 0; i < cnt; i++) {
		if (!isprint(bb_common_bufsiz1[i])
		 && !isspace(bb_common_bufsiz1[i])) {
			return 0;
		}
	}
#endif
	return 1;
}


/* dump accumulated "unified" diff changes */
static void dump_unified_vec(FILE * f1, FILE * f2)
{
	struct context_vec *cvp = context_vec_start;
	int lowa, upb, lowc, upd;
	int a, b, c, d;
	char ch;

	if (context_vec_start > context_vec_ptr)
		return;

	b = d = 0;			/* gcc */
	lowa = MAX(1, cvp->a - context);
	upb = MIN(len[0], context_vec_ptr->b + context);
	lowc = MAX(1, cvp->c - context);
	upd = MIN(len[1], context_vec_ptr->d + context);

	fputs("@@ -", stdout);
	uni_range(lowa, upb);
	fputs(" +", stdout);
	uni_range(lowc, upd);
	fputs(" @@", stdout);
	putchar('\n');

	/*
	 * Output changes in "unified" diff format--the old and new lines
	 * are printed together.
	 */
	for (; cvp <= context_vec_ptr; cvp++) {
		a = cvp->a;
		b = cvp->b;
		c = cvp->c;
		d = cvp->d;

		/*
		 * c: both new and old changes
		 * d: only changes in the old file
		 * a: only changes in the new file
		 */
		if (a <= b && c <= d)
			ch = 'c';
		else
			ch = (a <= b) ? 'd' : 'a';
		if (ch == 'c' || ch == 'd') {
			fetch(ixold, lowa, a - 1, f1, ' ');
			fetch(ixold, a, b, f1, '-');
		}
		if (ch == 'a')
			fetch(ixnew, lowc, c - 1, f2, ' ');
		if (ch == 'c' || ch == 'a')
			fetch(ixnew, c, d, f2, '+');
		lowa = b + 1;
		lowc = d + 1;
	}
	fetch(ixnew, d + 1, upd, f2, ' ');

	context_vec_ptr = context_vec_start - 1;
}


static void print_header(const char *file1, const char *file2)
{
	if (label[0] != NULL)
		printf("%s %s\n", "---", label[0]);
	else
		printf("%s %s\t%s", "---", file1, ctime(&stb1.st_mtime));
	if (label[1] != NULL)
		printf("%s %s\n", "+++", label[1]);
	else
		printf("%s %s\t%s", "+++", file2, ctime(&stb2.st_mtime));
}


/*
 * Indicate that there is a difference between lines a and b of the from file
 * to get to lines c to d of the to file.  If a is greater than b then there
 * are no lines in the from file involved and this means that there were
 * lines appended (beginning at b).  If c is greater than d then there are
 * lines missing from the to file.
 */
static void change(char *file1, FILE * f1, char *file2, FILE * f2, int a,
				   int b, int c, int d)
{
	static size_t max_context = 64;

	if ((a > b && c > d) || (option_mask32 & FLAG_q)) {
		anychange = 1;
		return;
	}

	/*
	 * Allocate change records as needed.
	 */
	if (context_vec_ptr == context_vec_end - 1) {
		ptrdiff_t offset = context_vec_ptr - context_vec_start;

		max_context <<= 1;
		context_vec_start = xrealloc(context_vec_start,
									 max_context *
									 sizeof(struct context_vec));
		context_vec_end = context_vec_start + max_context;
		context_vec_ptr = context_vec_start + offset;
	}
	if (anychange == 0) {
		/*
		 * Print the context/unidiff header first time through.
		 */
		print_header(file1, file2);
	} else if (a > context_vec_ptr->b + (2 * context) + 1 &&
			   c > context_vec_ptr->d + (2 * context) + 1) {
		/*
		 * If this change is more than 'context' lines from the
		 * previous change, dump the record and reset it.
		 */
		dump_unified_vec(f1, f2);
	}
	context_vec_ptr++;
	context_vec_ptr->a = a;
	context_vec_ptr->b = b;
	context_vec_ptr->c = c;
	context_vec_ptr->d = d;
	anychange = 1;
}


static void output(char *file1, FILE * f1, char *file2, FILE * f2)
{
	/* Note that j0 and j1 can't be used as they are defined in math.h.
	 * This also allows the rather amusing variable 'j00'... */
	int m, i0, i1, j00, j01;

	rewind(f1);
	rewind(f2);
	m = len[0];
	J[0] = 0;
	J[m + 1] = len[1] + 1;
	for (i0 = 1; i0 <= m; i0 = i1 + 1) {
		while (i0 <= m && J[i0] == J[i0 - 1] + 1)
			i0++;
		j00 = J[i0 - 1] + 1;
		i1 = i0 - 1;
		while (i1 < m && J[i1 + 1] == 0)
			i1++;
		j01 = J[i1 + 1] - 1;
		J[i1] = j01;
		change(file1, f1, file2, f2, i0, i1, j00, j01);
	}
	if (m == 0) {
		change(file1, f1, file2, f2, 1, 0, 1, len[1]);
	}
	if (anychange != 0 && !(option_mask32 & FLAG_q)) {
		dump_unified_vec(f1, f2);
	}
}

/*
 *      The following code uses an algorithm due to Harold Stone,
 *      which finds a pair of longest identical subsequences in
 *      the two files.
 *
 *      The major goal is to generate the match vector J.
 *      J[i] is the index of the line in file1 corresponding
 *      to line i file0. J[i] = 0 if there is no
 *      such line in file1.
 *
 *      Lines are hashed so as to work in core. All potential
 *      matches are located by sorting the lines of each file
 *      on the hash (called ``value''). In particular, this
 *      collects the equivalence classes in file1 together.
 *      Subroutine equiv replaces the value of each line in
 *      file0 by the index of the first element of its
 *      matching equivalence in (the reordered) file1.
 *      To save space equiv squeezes file1 into a single
 *      array member in which the equivalence classes
 *      are simply concatenated, except that their first
 *      members are flagged by changing sign.
 *
 *      Next the indices that point into member are unsorted into
 *      array class according to the original order of file0.
 *
 *      The cleverness lies in routine stone. This marches
 *      through the lines of file0, developing a vector klist
 *      of "k-candidates". At step i a k-candidate is a matched
 *      pair of lines x,y (x in file0 y in file1) such that
 *      there is a common subsequence of length k
 *      between the first i lines of file0 and the first y
 *      lines of file1, but there is no such subsequence for
 *      any smaller y. x is the earliest possible mate to y
 *      that occurs in such a subsequence.
 *
 *      Whenever any of the members of the equivalence class of
 *      lines in file1 matable to a line in file0 has serial number
 *      less than the y of some k-candidate, that k-candidate
 *      with the smallest such y is replaced. The new
 *      k-candidate is chained (via pred) to the current
 *      k-1 candidate so that the actual subsequence can
 *      be recovered. When a member has serial number greater
 *      that the y of all k-candidates, the klist is extended.
 *      At the end, the longest subsequence is pulled out
 *      and placed in the array J by unravel
 *
 *      With J in hand, the matches there recorded are
 *      checked against reality to assure that no spurious
 *      matches have crept in due to hashing. If they have,
 *      they are broken, and "jackpot" is recorded--a harmless
 *      matter except that a true match for a spuriously
 *      mated line may now be unnecessarily reported as a change.
 *
 *      Much of the complexity of the program comes simply
 *      from trying to minimize core utilization and
 *      maximize the range of doable problems by dynamically
 *      allocating what is needed and reusing what is not.
 *      The core requirements for problems larger than somewhat
 *      are (in words) 2*length(file0) + length(file1) +
 *      3*(number of k-candidates installed),  typically about
 *      6n words for files of length n.
 */
static int diffreg(char *ofile1, char *ofile2, int flags)
{
	char *file1 = ofile1;
	char *file2 = ofile2;
	FILE *f1 = NULL;
	FILE *f2 = NULL;
	int rval = D_SAME;
	int i;

	anychange = 0;
	context_vec_ptr = context_vec_start - 1;

	if (S_ISDIR(stb1.st_mode) != S_ISDIR(stb2.st_mode))
		return (S_ISDIR(stb1.st_mode) ? D_MISMATCH1 : D_MISMATCH2);
	if (LONE_DASH(file1) && LONE_DASH(file2))
		goto closem;

	f1 = stdin;
	if (flags & D_EMPTY1)
		f1 = xfopen(bb_dev_null, "r");
	else if (NOT_LONE_DASH(file1))
		f1 = xfopen(file1, "r");

	f2 = stdin;
	if (flags & D_EMPTY2)
		f2 = xfopen(bb_dev_null, "r");
	else if (NOT_LONE_DASH(file2))
		f2 = xfopen(file2, "r");

	i = files_differ(f1, f2, flags);
	if (i == 0)
		goto closem;
	else if (i != 1) {	/* 1 == ok */
		/* error */
		status |= 2;
		goto closem;
	}

	if (!asciifile(f1) || !asciifile(f2)) {
		rval = D_BINARY;
		status |= 1;
		goto closem;
	}

	prepare(0, f1, stb1.st_size);
	prepare(1, f2, stb2.st_size);
	prune();
	sort(sfile[0], slen[0]);
	sort(sfile[1], slen[1]);

	member = (int *) file[1];
	equiv(sfile[0], slen[0], sfile[1], slen[1], member);
	member = xrealloc(member, (slen[1] + 2) * sizeof(int));

	class = (int *) file[0];
	unsort(sfile[0], slen[0], class);
	class = xrealloc(class, (slen[0] + 2) * sizeof(int));

	klist = xmalloc((slen[0] + 2) * sizeof(int));
	clen = 0;
	clistlen = 100;
	clist = xmalloc(clistlen * sizeof(struct cand));
	i = stone(class, slen[0], member, klist);
	free(member);
	free(class);

	J = xrealloc(J, (len[0] + 2) * sizeof(int));
	unravel(klist[i]);
	free(clist);
	free(klist);

	ixold = xrealloc(ixold, (len[0] + 2) * sizeof(long));
	ixnew = xrealloc(ixnew, (len[1] + 2) * sizeof(long));
	check(f1, f2);
	output(file1, f1, file2, f2);

 closem:
	if (anychange) {
		status |= 1;
		if (rval == D_SAME)
			rval = D_DIFFER;
	}
	if (f1 != NULL)
		fclose(f1);
	if (f2 != NULL)
		fclose(f2);
	if (file1 != ofile1)
		free(file1);
	if (file2 != ofile2)
		free(file2);
	return rval;
}


#if ENABLE_FEATURE_DIFF_DIR
static void do_diff(char *dir1, char *path1, char *dir2, char *path2)
{
	int flags = D_HEADER;
	int val;

	char *fullpath1 = xasprintf("%s/%s", dir1, path1);
	char *fullpath2 = xasprintf("%s/%s", dir2, path2);

	if (stat(fullpath1, &stb1) != 0) {
		flags |= D_EMPTY1;
		memset(&stb1, 0, sizeof(stb1));
		fullpath1 = xasprintf("%s/%s", dir1, path2);
	}
	if (stat(fullpath2, &stb2) != 0) {
		flags |= D_EMPTY2;
		memset(&stb2, 0, sizeof(stb2));
		stb2.st_mode = stb1.st_mode;
		fullpath2 = xasprintf("%s/%s", dir2, path1);
	}

	if (stb1.st_mode == 0)
		stb1.st_mode = stb2.st_mode;

	if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) {
		printf("Common subdirectories: %s and %s\n", fullpath1, fullpath2);
		return;
	}

	if (!S_ISREG(stb1.st_mode) && !S_ISDIR(stb1.st_mode))
		val = D_SKIPPED1;
	else if (!S_ISREG(stb2.st_mode) && !S_ISDIR(stb2.st_mode))
		val = D_SKIPPED2;
	else
		val = diffreg(fullpath1, fullpath2, flags);

	print_status(val, fullpath1, fullpath2, NULL);
}
#endif


#if ENABLE_FEATURE_DIFF_DIR
static int dir_strcmp(const void *p1, const void *p2)
{
	return strcmp(*(char *const *) p1, *(char *const *) p2);
}


/* This function adds a filename to dl, the directory listing. */
static int add_to_dirlist(const char *filename,
		struct stat ATTRIBUTE_UNUSED * sb, void *userdata,
		int depth ATTRIBUTE_UNUSED)
{
	dl_count++;
	dl = xrealloc(dl, dl_count * sizeof(char *));
	dl[dl_count - 1] = xstrdup(filename);
	if (option_mask32 & FLAG_r) {
		int *pp = (int *) userdata;
		int path_len = *pp + 1;

		dl[dl_count - 1] = &(dl[dl_count - 1])[path_len];
	}
	return TRUE;
}


/* This returns a sorted directory listing. */
static char **get_dir(char *path)
{
	int i;
	char **retval;

	/* If -r has been set, then the recursive_action function will be
	 * used. Unfortunately, this outputs the root directory along with
	 * the recursed paths, so use void *userdata to specify the string
	 * length of the root directory. It can then be removed in
	 * add_to_dirlist. */

	int path_len = strlen(path);
	void *userdata = &path_len;

	/* Reset dl_count - there's no need to free dl as xrealloc does
	 * the job nicely. */
	dl_count = 0;

	/* Now fill dl with a listing. */
	if (option_mask32 & FLAG_r)
		recursive_action(path, TRUE, TRUE, FALSE, add_to_dirlist, NULL,
						 userdata, 0);
	else {
		DIR *dp;
		struct dirent *ep;

		dp = warn_opendir(path);
		while ((ep = readdir(dp))) {
			if (!strcmp(ep->d_name, "..") || LONE_CHAR(ep->d_name, '.'))
				continue;
			add_to_dirlist(ep->d_name, NULL, NULL, 0);
		}
		closedir(dp);
	}

	/* Sort dl alphabetically. */
	qsort(dl, dl_count, sizeof(char *), dir_strcmp);

	/* Copy dl so that we can return it. */
	retval = xmalloc(dl_count * sizeof(char *));
	for (i = 0; i < dl_count; i++)
		retval[i] = xstrdup(dl[i]);

	return retval;
}


static void diffdir(char *p1, char *p2)
{
	char **dirlist1, **dirlist2;
	char *dp1, *dp2;
	int dirlist1_count, dirlist2_count;
	int pos;

	/* Check for trailing slashes. */

	dp1 = last_char_is(p1, '/');
	if (dp1 != NULL)
		*dp1 = '\0';
	dp2 = last_char_is(p2, '/');
	if (dp2 != NULL)
		*dp2 = '\0';

	/* Get directory listings for p1 and p2. */

	dirlist1 = get_dir(p1);
	dirlist1_count = dl_count;
	dirlist1[dirlist1_count] = NULL;
	dirlist2 = get_dir(p2);
	dirlist2_count = dl_count;
	dirlist2[dirlist2_count] = NULL;

	/* If -S was set, find the starting point. */
	if (start) {
		while (*dirlist1 != NULL && strcmp(*dirlist1, start) < 0)
			dirlist1++;
		while (*dirlist2 != NULL && strcmp(*dirlist2, start) < 0)
			dirlist2++;
		if ((*dirlist1 == NULL) || (*dirlist2 == NULL))
			bb_error_msg(bb_msg_invalid_arg, "NULL", "-S");
	}

	/* Now that both dirlist1 and dirlist2 contain sorted directory
	 * listings, we can start to go through dirlist1. If both listings
	 * contain the same file, then do a normal diff. Otherwise, behaviour
	 * is determined by whether the -N flag is set. */
	while (*dirlist1 != NULL || *dirlist2 != NULL) {
		dp1 = *dirlist1;
		dp2 = *dirlist2;
		pos = dp1 == NULL ? 1 : dp2 == NULL ? -1 : strcmp(dp1, dp2);
		if (pos == 0) {
			do_diff(p1, dp1, p2, dp2);
			dirlist1++;
			dirlist2++;
		} else if (pos < 0) {
			if (option_mask32 & FLAG_N)
				do_diff(p1, dp1, p2, NULL);
			else
				print_only(p1, strlen(p1) + 1, dp1);
			dirlist1++;
		} else {
			if (option_mask32 & FLAG_N)
				do_diff(p1, NULL, p2, dp2);
			else
				print_only(p2, strlen(p2) + 1, dp2);
			dirlist2++;
		}
	}
}
#endif


int diff_main(int argc, char **argv)
{
	int gotstdin = 0;

	char *U_opt;
	llist_t *L_arg = NULL;

	opt_complementary = "L::";
	getopt32(argc, argv, "abdiL:NqrsS:tTU:wu"
			"p" /* ignored (for compatibility) */,
			&L_arg, &start, &U_opt);

	if (option_mask32 & FLAG_L) {
		while (L_arg) {
			if (label[0] == NULL)
				label[0] = L_arg->data;
			else if (label[1] == NULL)
				label[1] = L_arg->data;
			else
				bb_show_usage();

			L_arg = L_arg->link;
		}

		/* If both label[0] and label[1] were set, they need to be swapped. */
		if (label[0] && label[1]) {
			char *tmp;

			tmp = label[1];
			label[1] = label[0];
			label[0] = tmp;
		}
	}

	context = 3;		/* This is the default number of lines of context. */
	if (option_mask32 & FLAG_U) {
		context = xatou_range(U_opt, 1, INT_MAX);
	}
	argc -= optind;
	argv += optind;

	/*
	 * Do sanity checks, fill in stb1 and stb2 and call the appropriate
	 * driver routine.  Both drivers use the contents of stb1 and stb2.
	 */
	if (argc < 2) {
		bb_error_msg("missing filename");
		bb_show_usage();
	}
	if (LONE_DASH(argv[0])) {
		fstat(STDIN_FILENO, &stb1);
		gotstdin = 1;
	} else
		xstat(argv[0], &stb1);
	if (LONE_DASH(argv[1])) {
		fstat(STDIN_FILENO, &stb2);
		gotstdin = 1;
	} else
		xstat(argv[1], &stb2);
	if (gotstdin && (S_ISDIR(stb1.st_mode) || S_ISDIR(stb2.st_mode)))
		bb_error_msg_and_die("can't compare - to a directory");
	if (S_ISDIR(stb1.st_mode) && S_ISDIR(stb2.st_mode)) {
#if ENABLE_FEATURE_DIFF_DIR
		diffdir(argv[0], argv[1]);
#else
		bb_error_msg_and_die("directory comparison not supported");
#endif
	} else {
		if (S_ISDIR(stb1.st_mode)) {
			argv[0] = concat_path_file(argv[0], argv[1]);
			xstat(argv[0], &stb1);
		}
		if (S_ISDIR(stb2.st_mode)) {
			argv[1] = concat_path_file(argv[1], argv[0]);
			xstat(argv[1], &stb2);
		}
		print_status(diffreg(argv[0], argv[1], 0), argv[0], argv[1], NULL);
	}
	return status;
}