xref: /freebsd/usr.bin/diff/diffreg.c (revision 02e9120893770924227138ba49df1edb3896112a)
1 /*	$OpenBSD: diffreg.c,v 1.93 2019/06/28 13:35:00 deraadt Exp $	*/
2 
3 /*-
4  * SPDX-License-Identifier: BSD-4-Clause
5  *
6  * Copyright (C) Caldera International Inc.  2001-2002.
7  * All rights reserved.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  * 1. Redistributions of source code and documentation must retain the above
13  *    copyright notice, this list of conditions and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  * 3. All advertising materials mentioning features or use of this software
18  *    must display the following acknowledgement:
19  *	This product includes software developed or owned by Caldera
20  *	International, Inc.
21  * 4. Neither the name of Caldera International, Inc. nor the names of other
22  *    contributors may be used to endorse or promote products derived from
23  *    this software without specific prior written permission.
24  *
25  * USE OF THE SOFTWARE PROVIDED FOR UNDER THIS LICENSE BY CALDERA
26  * INTERNATIONAL, INC. AND CONTRIBUTORS ``AS IS'' AND ANY EXPRESS OR
27  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
28  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
29  * IN NO EVENT SHALL CALDERA INTERNATIONAL, INC. BE LIABLE FOR ANY DIRECT,
30  * INDIRECT INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
31  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
32  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
33  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
34  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
35  * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
36  * POSSIBILITY OF SUCH DAMAGE.
37  */
38 /*-
39  * Copyright (c) 1991, 1993
40  *	The Regents of the University of California.  All rights reserved.
41  *
42  * Redistribution and use in source and binary forms, with or without
43  * modification, are permitted provided that the following conditions
44  * are met:
45  * 1. Redistributions of source code must retain the above copyright
46  *    notice, this list of conditions and the following disclaimer.
47  * 2. Redistributions in binary form must reproduce the above copyright
48  *    notice, this list of conditions and the following disclaimer in the
49  *    documentation and/or other materials provided with the distribution.
50  * 3. Neither the name of the University nor the names of its contributors
51  *    may be used to endorse or promote products derived from this software
52  *    without specific prior written permission.
53  *
54  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
55  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
56  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
57  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
58  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
59  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
60  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
61  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
62  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
63  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
64  * SUCH DAMAGE.
65  */
66 
67 #include <sys/capsicum.h>
68 #include <sys/stat.h>
69 
70 #include <capsicum_helpers.h>
71 #include <ctype.h>
72 #include <err.h>
73 #include <errno.h>
74 #include <fcntl.h>
75 #include <math.h>
76 #include <paths.h>
77 #include <regex.h>
78 #include <stdbool.h>
79 #include <stddef.h>
80 #include <stdint.h>
81 #include <stdio.h>
82 #include <stdlib.h>
83 #include <string.h>
84 
85 #include "pr.h"
86 #include "diff.h"
87 #include "xmalloc.h"
88 
89 /*
90  * diff - compare two files.
91  */
92 
93 /*
94  *	Uses an algorithm due to Harold Stone, which finds a pair of longest
95  *	identical subsequences in the two files.
96  *
97  *	The major goal is to generate the match vector J. J[i] is the index of
98  *	the line in file1 corresponding to line i file0. J[i] = 0 if there is no
99  *	such line in file1.
100  *
101  *	Lines are hashed so as to work in core. All potential matches are
102  *	located by sorting the lines of each file on the hash (called
103  *	``value''). In particular, this collects the equivalence classes in
104  *	file1 together. Subroutine equiv replaces the value of each line in
105  *	file0 by the index of the first element of its matching equivalence in
106  *	(the reordered) file1. To save space equiv squeezes file1 into a single
107  *	array member in which the equivalence classes are simply concatenated,
108  *	except that their first members are flagged by changing sign.
109  *
110  *	Next the indices that point into member are unsorted into array class
111  *	according to the original order of file0.
112  *
113  *	The cleverness lies in routine stone. This marches through the lines of
114  *	file0, developing a vector klist of "k-candidates". At step i
115  *	a k-candidate is a matched pair of lines x,y (x in file0 y in file1)
116  *	such that there is a common subsequence of length k between the first
117  *	i lines of file0 and the first y lines of file1, but there is no such
118  *	subsequence for any smaller y. x is the earliest possible mate to y that
119  *	occurs in such a subsequence.
120  *
121  *	Whenever any of the members of the equivalence class of lines in file1
122  *	matable to a line in file0 has serial number less than the y of some
123  *	k-candidate, that k-candidate with the smallest such y is replaced. The
124  *	new k-candidate is chained (via pred) to the current k-1 candidate so
125  *	that the actual subsequence can be recovered. When a member has serial
126  *	number greater that the y of all k-candidates, the klist is extended. At
127  *	the end, the longest subsequence is pulled out and placed in the array J
128  *	by unravel.
129  *
130  *	With J in hand, the matches there recorded are check'ed against reality
131  *	to assure that no spurious matches have crept in due to hashing. If they
132  *	have, they are broken, and "jackpot" is recorded -- a harmless matter
133  *	except that a true match for a spuriously mated line may now be
134  *	unnecessarily reported as a change.
135  *
136  *	Much of the complexity of the program comes simply from trying to
137  *	minimize core utilization and maximize the range of doable problems by
138  *	dynamically allocating what is needed and reusing what is not. The core
139  *	requirements for problems larger than somewhat are (in words)
140  *	2*length(file0) + length(file1) + 3*(number of k-candidates installed),
141  *	typically about 6n words for files of length n.
142  */
143 
144 struct cand {
145 	int	x;
146 	int	y;
147 	int	pred;
148 };
149 
150 static struct line {
151 	int	serial;
152 	int	value;
153 } *file[2];
154 
155 /*
156  * The following struct is used to record change information when
157  * doing a "context" or "unified" diff.  (see routine "change" to
158  * understand the highly mnemonic field names)
159  */
160 struct context_vec {
161 	int	a;		/* start line in old file */
162 	int	b;		/* end line in old file */
163 	int	c;		/* start line in new file */
164 	int	d;		/* end line in new file */
165 };
166 
167 enum readhash { RH_BINARY, RH_OK, RH_EOF };
168 
169 #define MIN_PAD		1
170 static FILE	*opentemp(const char *);
171 static void	 output(char *, FILE *, char *, FILE *, int);
172 static void	 check(FILE *, FILE *, int);
173 static void	 range(int, int, const char *);
174 static void	 uni_range(int, int);
175 static void	 dump_context_vec(FILE *, FILE *, int);
176 static void	 dump_unified_vec(FILE *, FILE *, int);
177 static bool	 prepare(int, FILE *, size_t, int);
178 static void	 prune(void);
179 static void	 equiv(struct line *, int, struct line *, int, int *);
180 static void	 unravel(int);
181 static void	 unsort(struct line *, int, int *);
182 static void	 change(char *, FILE *, char *, FILE *, int, int, int, int, int *);
183 static void	 sort(struct line *, int);
184 static void	 print_header(const char *, const char *);
185 static void	 print_space(int, int, int);
186 static bool	 ignoreline_pattern(char *);
187 static bool	 ignoreline(char *, bool);
188 static int	 asciifile(FILE *);
189 static int	 fetch(long *, int, int, FILE *, int, int, int);
190 static int	 newcand(int, int, int);
191 static int	 search(int *, int, int);
192 static int	 skipline(FILE *);
193 static int	 stone(int *, int, int *, int *, int);
194 static enum readhash readhash(FILE *, int, unsigned *);
195 static int	 files_differ(FILE *, FILE *, int);
196 static char	*match_function(const long *, int, FILE *);
197 static char	*preadline(int, size_t, off_t);
198 
199 static int	 *J;			/* will be overlaid on class */
200 static int	 *class;		/* will be overlaid on file[0] */
201 static int	 *klist;		/* will be overlaid on file[0] after class */
202 static int	 *member;		/* will be overlaid on file[1] */
203 static int	 clen;
204 static int	 inifdef;		/* whether or not we are in a #ifdef block */
205 static int	 len[2];
206 static int	 pref, suff;	/* length of prefix and suffix */
207 static int	 slen[2];
208 static int	 anychange;
209 static int	 hw, padding;	/* half width and padding */
210 static int	 edoffset;
211 static long	*ixnew;		/* will be overlaid on file[1] */
212 static long	*ixold;		/* will be overlaid on klist */
213 static struct cand *clist;	/* merely a free storage pot for candidates */
214 static int	 clistlen;		/* the length of clist */
215 static struct line *sfile[2];	/* shortened by pruning common prefix/suffix */
216 static int	(*chrtran)(int);	/* translation table for case-folding */
217 static struct context_vec *context_vec_start;
218 static struct context_vec *context_vec_end;
219 static struct context_vec *context_vec_ptr;
220 
221 #define FUNCTION_CONTEXT_SIZE	55
222 static char lastbuf[FUNCTION_CONTEXT_SIZE];
223 static int lastline;
224 static int lastmatchline;
225 
226 static int
227 clow2low(int c)
228 {
229 
230 	return (c);
231 }
232 
233 static int
234 cup2low(int c)
235 {
236 
237 	return (tolower(c));
238 }
239 
240 int
241 diffreg(char *file1, char *file2, int flags, int capsicum)
242 {
243 	FILE *f1, *f2;
244 	int i, rval;
245 	struct pr *pr = NULL;
246 	cap_rights_t rights_ro;
247 
248 	f1 = f2 = NULL;
249 	rval = D_SAME;
250 	anychange = 0;
251 	lastline = 0;
252 	lastmatchline = 0;
253 
254 	 /*
255 	  * hw excludes padding and make sure when -t is not used,
256 	  * the second column always starts from the closest tab stop
257 	  */
258 	if (diff_format == D_SIDEBYSIDE) {
259 		hw = width >> 1;
260 		padding = tabsize - (hw % tabsize);
261 		if ((flags & D_EXPANDTABS) != 0 || (padding % tabsize == 0))
262 			padding = MIN_PAD;
263 
264 		hw = (width >> 1) -
265 		    ((padding == MIN_PAD) ? (padding << 1) : padding) - 1;
266 	}
267 
268 
269 	if (flags & D_IGNORECASE)
270 		chrtran = cup2low;
271 	else
272 		chrtran = clow2low;
273 	if (S_ISDIR(stb1.st_mode) != S_ISDIR(stb2.st_mode))
274 		return (S_ISDIR(stb1.st_mode) ? D_MISMATCH1 : D_MISMATCH2);
275 	if (strcmp(file1, "-") == 0 && strcmp(file2, "-") == 0)
276 		goto closem;
277 
278 	if (flags & D_EMPTY1)
279 		f1 = fopen(_PATH_DEVNULL, "r");
280 	else {
281 		if (!S_ISREG(stb1.st_mode)) {
282 			if ((f1 = opentemp(file1)) == NULL ||
283 			    fstat(fileno(f1), &stb1) == -1) {
284 				warn("%s", file1);
285 				rval = D_ERROR;
286 				status |= 2;
287 				goto closem;
288 			}
289 		} else if (strcmp(file1, "-") == 0)
290 			f1 = stdin;
291 		else
292 			f1 = fopen(file1, "r");
293 	}
294 	if (f1 == NULL) {
295 		warn("%s", file1);
296 		rval = D_ERROR;
297 		status |= 2;
298 		goto closem;
299 	}
300 
301 	if (flags & D_EMPTY2)
302 		f2 = fopen(_PATH_DEVNULL, "r");
303 	else {
304 		if (!S_ISREG(stb2.st_mode)) {
305 			if ((f2 = opentemp(file2)) == NULL ||
306 			    fstat(fileno(f2), &stb2) == -1) {
307 				warn("%s", file2);
308 				rval = D_ERROR;
309 				status |= 2;
310 				goto closem;
311 			}
312 		} else if (strcmp(file2, "-") == 0)
313 			f2 = stdin;
314 		else
315 			f2 = fopen(file2, "r");
316 	}
317 	if (f2 == NULL) {
318 		warn("%s", file2);
319 		rval = D_ERROR;
320 		status |= 2;
321 		goto closem;
322 	}
323 
324 	if (lflag)
325 		pr = start_pr(file1, file2);
326 
327 	if (capsicum) {
328 		cap_rights_init(&rights_ro, CAP_READ, CAP_FSTAT, CAP_SEEK);
329 		if (caph_rights_limit(fileno(f1), &rights_ro) < 0)
330 			err(2, "unable to limit rights on: %s", file1);
331 		if (caph_rights_limit(fileno(f2), &rights_ro) < 0)
332 			err(2, "unable to limit rights on: %s", file2);
333 		if (fileno(f1) == STDIN_FILENO || fileno(f2) == STDIN_FILENO) {
334 			/* stdin has already been limited */
335 			if (caph_limit_stderr() == -1)
336 				err(2, "unable to limit stderr");
337 			if (caph_limit_stdout() == -1)
338 				err(2, "unable to limit stdout");
339 		} else if (caph_limit_stdio() == -1)
340 				err(2, "unable to limit stdio");
341 
342 		caph_cache_catpages();
343 		caph_cache_tzdata();
344 		if (caph_enter() < 0)
345 			err(2, "unable to enter capability mode");
346 	}
347 
348 	switch (files_differ(f1, f2, flags)) {
349 	case 0:
350 		goto closem;
351 	case 1:
352 		break;
353 	default:
354 		/* error */
355 		rval = D_ERROR;
356 		status |= 2;
357 		goto closem;
358 	}
359 
360 	if (diff_format == D_BRIEF && ignore_pats == NULL &&
361 	    (flags & (D_FOLDBLANKS|D_IGNOREBLANKS|D_IGNORECASE|D_STRIPCR)) == 0)
362 	{
363 		rval = D_DIFFER;
364 		status |= 1;
365 		goto closem;
366 	}
367 	if ((flags & D_FORCEASCII) != 0) {
368 		(void)prepare(0, f1, stb1.st_size, flags);
369 		(void)prepare(1, f2, stb2.st_size, flags);
370 	} else if (!asciifile(f1) || !asciifile(f2) ||
371 		    !prepare(0, f1, stb1.st_size, flags) ||
372 		    !prepare(1, f2, stb2.st_size, flags)) {
373 		rval = D_BINARY;
374 		status |= 1;
375 		goto closem;
376 	}
377 
378 	prune();
379 	sort(sfile[0], slen[0]);
380 	sort(sfile[1], slen[1]);
381 
382 	member = (int *)file[1];
383 	equiv(sfile[0], slen[0], sfile[1], slen[1], member);
384 	member = xreallocarray(member, slen[1] + 2, sizeof(*member));
385 
386 	class = (int *)file[0];
387 	unsort(sfile[0], slen[0], class);
388 	class = xreallocarray(class, slen[0] + 2, sizeof(*class));
389 
390 	klist = xcalloc(slen[0] + 2, sizeof(*klist));
391 	clen = 0;
392 	clistlen = 100;
393 	clist = xcalloc(clistlen, sizeof(*clist));
394 	i = stone(class, slen[0], member, klist, flags);
395 	free(member);
396 	free(class);
397 
398 	J = xreallocarray(J, len[0] + 2, sizeof(*J));
399 	unravel(klist[i]);
400 	free(clist);
401 	free(klist);
402 
403 	ixold = xreallocarray(ixold, len[0] + 2, sizeof(*ixold));
404 	ixnew = xreallocarray(ixnew, len[1] + 2, sizeof(*ixnew));
405 	check(f1, f2, flags);
406 	output(file1, f1, file2, f2, flags);
407 
408 closem:
409 	if (pr != NULL)
410 		stop_pr(pr);
411 	if (anychange) {
412 		status |= 1;
413 		if (rval == D_SAME)
414 			rval = D_DIFFER;
415 	}
416 	if (f1 != NULL)
417 		fclose(f1);
418 	if (f2 != NULL)
419 		fclose(f2);
420 
421 	return (rval);
422 }
423 
424 /*
425  * Check to see if the given files differ.
426  * Returns 0 if they are the same, 1 if different, and -1 on error.
427  * XXX - could use code from cmp(1) [faster]
428  */
429 static int
430 files_differ(FILE *f1, FILE *f2, int flags)
431 {
432 	char buf1[BUFSIZ], buf2[BUFSIZ];
433 	size_t i, j;
434 
435 	if ((flags & (D_EMPTY1|D_EMPTY2)) || stb1.st_size != stb2.st_size ||
436 	    (stb1.st_mode & S_IFMT) != (stb2.st_mode & S_IFMT))
437 		return (1);
438 
439 	if (stb1.st_dev == stb2.st_dev && stb1.st_ino == stb2.st_ino)
440 		return (0);
441 
442 	for (;;) {
443 		i = fread(buf1, 1, sizeof(buf1), f1);
444 		j = fread(buf2, 1, sizeof(buf2), f2);
445 		if ((!i && ferror(f1)) || (!j && ferror(f2)))
446 			return (-1);
447 		if (i != j)
448 			return (1);
449 		if (i == 0)
450 			return (0);
451 		if (memcmp(buf1, buf2, i) != 0)
452 			return (1);
453 	}
454 }
455 
456 static FILE *
457 opentemp(const char *f)
458 {
459 	char buf[BUFSIZ], tempfile[PATH_MAX];
460 	ssize_t nread;
461 	int ifd, ofd;
462 
463 	if (strcmp(f, "-") == 0)
464 		ifd = STDIN_FILENO;
465 	else if ((ifd = open(f, O_RDONLY, 0644)) == -1)
466 		return (NULL);
467 
468 	(void)strlcpy(tempfile, _PATH_TMP "/diff.XXXXXXXX", sizeof(tempfile));
469 
470 	if ((ofd = mkstemp(tempfile)) == -1) {
471 		close(ifd);
472 		return (NULL);
473 	}
474 	unlink(tempfile);
475 	while ((nread = read(ifd, buf, BUFSIZ)) > 0) {
476 		if (write(ofd, buf, nread) != nread) {
477 			close(ifd);
478 			close(ofd);
479 			return (NULL);
480 		}
481 	}
482 	close(ifd);
483 	lseek(ofd, (off_t)0, SEEK_SET);
484 	return (fdopen(ofd, "r"));
485 }
486 
487 static bool
488 prepare(int i, FILE *fd, size_t filesize, int flags)
489 {
490 	struct line *p;
491 	unsigned h;
492 	size_t sz, j = 0;
493 	enum readhash r;
494 
495 	rewind(fd);
496 
497 	sz = MIN(filesize, SIZE_MAX) / 25;
498 	if (sz < 100)
499 		sz = 100;
500 
501 	p = xcalloc(sz + 3, sizeof(*p));
502 	while ((r = readhash(fd, flags, &h)) != RH_EOF)
503 		switch (r) {
504 		case RH_EOF: /* otherwise clang complains */
505 		case RH_BINARY:
506 			return (false);
507 		case RH_OK:
508 			if (j == sz) {
509 				sz = sz * 3 / 2;
510 				p = xreallocarray(p, sz + 3, sizeof(*p));
511 			}
512 			p[++j].value = h;
513 		}
514 
515 	len[i] = j;
516 	file[i] = p;
517 
518 	return (true);
519 }
520 
521 static void
522 prune(void)
523 {
524 	int i, j;
525 
526 	for (pref = 0; pref < len[0] && pref < len[1] &&
527 	    file[0][pref + 1].value == file[1][pref + 1].value;
528 	    pref++)
529 		;
530 	for (suff = 0; suff < len[0] - pref && suff < len[1] - pref &&
531 	    file[0][len[0] - suff].value == file[1][len[1] - suff].value;
532 	    suff++)
533 		;
534 	for (j = 0; j < 2; j++) {
535 		sfile[j] = file[j] + pref;
536 		slen[j] = len[j] - pref - suff;
537 		for (i = 0; i <= slen[j]; i++)
538 			sfile[j][i].serial = i;
539 	}
540 }
541 
542 static void
543 equiv(struct line *a, int n, struct line *b, int m, int *c)
544 {
545 	int i, j;
546 
547 	i = j = 1;
548 	while (i <= n && j <= m) {
549 		if (a[i].value < b[j].value)
550 			a[i++].value = 0;
551 		else if (a[i].value == b[j].value)
552 			a[i++].value = j;
553 		else
554 			j++;
555 	}
556 	while (i <= n)
557 		a[i++].value = 0;
558 	b[m + 1].value = 0;
559 	j = 0;
560 	while (++j <= m) {
561 		c[j] = -b[j].serial;
562 		while (b[j + 1].value == b[j].value) {
563 			j++;
564 			c[j] = b[j].serial;
565 		}
566 	}
567 	c[j] = -1;
568 }
569 
570 static int
571 stone(int *a, int n, int *b, int *c, int flags)
572 {
573 	int i, k, y, j, l;
574 	int oldc, tc, oldl, sq;
575 	unsigned numtries, bound;
576 
577 	if (flags & D_MINIMAL)
578 		bound = UINT_MAX;
579 	else {
580 		sq = sqrt(n);
581 		bound = MAX(256, sq);
582 	}
583 
584 	k = 0;
585 	c[0] = newcand(0, 0, 0);
586 	for (i = 1; i <= n; i++) {
587 		j = a[i];
588 		if (j == 0)
589 			continue;
590 		y = -b[j];
591 		oldl = 0;
592 		oldc = c[0];
593 		numtries = 0;
594 		do {
595 			if (y <= clist[oldc].y)
596 				continue;
597 			l = search(c, k, y);
598 			if (l != oldl + 1)
599 				oldc = c[l - 1];
600 			if (l <= k) {
601 				if (clist[c[l]].y <= y)
602 					continue;
603 				tc = c[l];
604 				c[l] = newcand(i, y, oldc);
605 				oldc = tc;
606 				oldl = l;
607 				numtries++;
608 			} else {
609 				c[l] = newcand(i, y, oldc);
610 				k++;
611 				break;
612 			}
613 		} while ((y = b[++j]) > 0 && numtries < bound);
614 	}
615 	return (k);
616 }
617 
618 static int
619 newcand(int x, int y, int pred)
620 {
621 	struct cand *q;
622 
623 	if (clen == clistlen) {
624 		clistlen = clistlen * 11 / 10;
625 		clist = xreallocarray(clist, clistlen, sizeof(*clist));
626 	}
627 	q = clist + clen;
628 	q->x = x;
629 	q->y = y;
630 	q->pred = pred;
631 	return (clen++);
632 }
633 
634 static int
635 search(int *c, int k, int y)
636 {
637 	int i, j, l, t;
638 
639 	if (clist[c[k]].y < y)	/* quick look for typical case */
640 		return (k + 1);
641 	i = 0;
642 	j = k + 1;
643 	for (;;) {
644 		l = (i + j) / 2;
645 		if (l <= i)
646 			break;
647 		t = clist[c[l]].y;
648 		if (t > y)
649 			j = l;
650 		else if (t < y)
651 			i = l;
652 		else
653 			return (l);
654 	}
655 	return (l + 1);
656 }
657 
658 static void
659 unravel(int p)
660 {
661 	struct cand *q;
662 	int i;
663 
664 	for (i = 0; i <= len[0]; i++)
665 		J[i] = i <= pref ? i :
666 		    i > len[0] - suff ? i + len[1] - len[0] : 0;
667 	for (q = clist + p; q->y != 0; q = clist + q->pred)
668 		J[q->x + pref] = q->y + pref;
669 }
670 
671 /*
672  * Check does double duty:
673  *  1. ferret out any fortuitous correspondences due to confounding by
674  *     hashing (which result in "jackpot")
675  *  2. collect random access indexes to the two files
676  */
677 static void
678 check(FILE *f1, FILE *f2, int flags)
679 {
680 	int i, j, /* jackpot, */ c, d;
681 	long ctold, ctnew;
682 
683 	rewind(f1);
684 	rewind(f2);
685 	j = 1;
686 	ixold[0] = ixnew[0] = 0;
687 	/* jackpot = 0; */
688 	ctold = ctnew = 0;
689 	for (i = 1; i <= len[0]; i++) {
690 		if (J[i] == 0) {
691 			ixold[i] = ctold += skipline(f1);
692 			continue;
693 		}
694 		while (j < J[i]) {
695 			ixnew[j] = ctnew += skipline(f2);
696 			j++;
697 		}
698 		if (flags & (D_FOLDBLANKS | D_IGNOREBLANKS | D_IGNORECASE | D_STRIPCR)) {
699 			for (;;) {
700 				c = getc(f1);
701 				d = getc(f2);
702 				/*
703 				 * GNU diff ignores a missing newline
704 				 * in one file for -b or -w.
705 				 */
706 				if (flags & (D_FOLDBLANKS | D_IGNOREBLANKS)) {
707 					if (c == EOF && d == '\n') {
708 						ctnew++;
709 						break;
710 					} else if (c == '\n' && d == EOF) {
711 						ctold++;
712 						break;
713 					}
714 				}
715 				ctold++;
716 				ctnew++;
717 				if (flags & D_STRIPCR && (c == '\r' || d == '\r')) {
718 					if (c == '\r') {
719 						if ((c = getc(f1)) == '\n') {
720 							ctold++;
721 						} else {
722 							ungetc(c, f1);
723 						}
724 					}
725 					if (d == '\r') {
726 						if ((d = getc(f2)) == '\n') {
727 							ctnew++;
728 						} else {
729 							ungetc(d, f2);
730 						}
731 					}
732 					break;
733 				}
734 				if ((flags & D_FOLDBLANKS) && isspace(c) &&
735 				    isspace(d)) {
736 					do {
737 						if (c == '\n')
738 							break;
739 						ctold++;
740 					} while (isspace(c = getc(f1)));
741 					do {
742 						if (d == '\n')
743 							break;
744 						ctnew++;
745 					} while (isspace(d = getc(f2)));
746 				} else if (flags & D_IGNOREBLANKS) {
747 					while (isspace(c) && c != '\n') {
748 						c = getc(f1);
749 						ctold++;
750 					}
751 					while (isspace(d) && d != '\n') {
752 						d = getc(f2);
753 						ctnew++;
754 					}
755 				}
756 				if (chrtran(c) != chrtran(d)) {
757 					/* jackpot++; */
758 					J[i] = 0;
759 					if (c != '\n' && c != EOF)
760 						ctold += skipline(f1);
761 					if (d != '\n' && c != EOF)
762 						ctnew += skipline(f2);
763 					break;
764 				}
765 				if (c == '\n' || c == EOF)
766 					break;
767 			}
768 		} else {
769 			for (;;) {
770 				ctold++;
771 				ctnew++;
772 				if ((c = getc(f1)) != (d = getc(f2))) {
773 					/* jackpot++; */
774 					J[i] = 0;
775 					if (c != '\n' && c != EOF)
776 						ctold += skipline(f1);
777 					if (d != '\n' && c != EOF)
778 						ctnew += skipline(f2);
779 					break;
780 				}
781 				if (c == '\n' || c == EOF)
782 					break;
783 			}
784 		}
785 		ixold[i] = ctold;
786 		ixnew[j] = ctnew;
787 		j++;
788 	}
789 	for (; j <= len[1]; j++) {
790 		ixnew[j] = ctnew += skipline(f2);
791 	}
792 	/*
793 	 * if (jackpot)
794 	 *	fprintf(stderr, "jackpot\n");
795 	 */
796 }
797 
798 /* shellsort CACM #201 */
799 static void
800 sort(struct line *a, int n)
801 {
802 	struct line *ai, *aim, w;
803 	int j, m = 0, k;
804 
805 	if (n == 0)
806 		return;
807 	for (j = 1; j <= n; j *= 2)
808 		m = 2 * j - 1;
809 	for (m /= 2; m != 0; m /= 2) {
810 		k = n - m;
811 		for (j = 1; j <= k; j++) {
812 			for (ai = &a[j]; ai > a; ai -= m) {
813 				aim = &ai[m];
814 				if (aim < ai)
815 					break;	/* wraparound */
816 				if (aim->value > ai[0].value ||
817 				    (aim->value == ai[0].value &&
818 					aim->serial > ai[0].serial))
819 					break;
820 				w.value = ai[0].value;
821 				ai[0].value = aim->value;
822 				aim->value = w.value;
823 				w.serial = ai[0].serial;
824 				ai[0].serial = aim->serial;
825 				aim->serial = w.serial;
826 			}
827 		}
828 	}
829 }
830 
831 static void
832 unsort(struct line *f, int l, int *b)
833 {
834 	int *a, i;
835 
836 	a = xcalloc(l + 1, sizeof(*a));
837 	for (i = 1; i <= l; i++)
838 		a[f[i].serial] = f[i].value;
839 	for (i = 1; i <= l; i++)
840 		b[i] = a[i];
841 	free(a);
842 }
843 
844 static int
845 skipline(FILE *f)
846 {
847 	int i, c;
848 
849 	for (i = 1; (c = getc(f)) != '\n' && c != EOF; i++)
850 		continue;
851 	return (i);
852 }
853 
854 static void
855 output(char *file1, FILE *f1, char *file2, FILE *f2, int flags)
856 {
857 	int i, j, m, i0, i1, j0, j1, nc;
858 
859 	rewind(f1);
860 	rewind(f2);
861 	m = len[0];
862 	J[0] = 0;
863 	J[m + 1] = len[1] + 1;
864 	if (diff_format != D_EDIT) {
865 		for (i0 = 1; i0 <= m; i0 = i1 + 1) {
866 			while (i0 <= m && J[i0] == J[i0 - 1] + 1) {
867 				if (diff_format == D_SIDEBYSIDE && suppress_common != 1) {
868 					nc = fetch(ixold, i0, i0, f1, '\0', 1, flags);
869 					print_space(nc, (hw - nc) + (padding << 1) + 1, flags);
870 					fetch(ixnew, J[i0], J[i0], f2, '\0', 0, flags);
871 					printf("\n");
872 				}
873 				i0++;
874 			}
875 			j0 = J[i0 - 1] + 1;
876 			i1 = i0 - 1;
877 			while (i1 < m && J[i1 + 1] == 0)
878 				i1++;
879 			j1 = J[i1 + 1] - 1;
880 			J[i1] = j1;
881 
882 			/*
883 			 * When using side-by-side, lines from both of the files are
884 			 * printed. The algorithm used by diff(1) identifies the ranges
885 			 * in which two files differ.
886 			 * See the change() function below.
887 			 * The for loop below consumes the shorter range, whereas one of
888 			 * the while loops deals with the longer one.
889 			 */
890 			if (diff_format == D_SIDEBYSIDE) {
891 				for (i = i0, j = j0; i <= i1 && j <= j1; i++, j++)
892 					change(file1, f1, file2, f2, i, i, j, j, &flags);
893 
894 				while (i <= i1) {
895 					change(file1, f1, file2, f2, i, i, j + 1, j, &flags);
896 					i++;
897 				}
898 
899 				while (j <= j1) {
900 					change(file1, f1, file2, f2, i + 1, i, j, j, &flags);
901 					j++;
902 				}
903 			} else
904 				change(file1, f1, file2, f2, i0, i1, j0, j1, &flags);
905 		}
906 	} else {
907 		for (i0 = m; i0 >= 1; i0 = i1 - 1) {
908 			while (i0 >= 1 && J[i0] == J[i0 + 1] - 1 && J[i0] != 0)
909 				i0--;
910 			j0 = J[i0 + 1] - 1;
911 			i1 = i0 + 1;
912 			while (i1 > 1 && J[i1 - 1] == 0)
913 				i1--;
914 			j1 = J[i1 - 1] + 1;
915 			J[i1] = j1;
916 			change(file1, f1, file2, f2, i1, i0, j1, j0, &flags);
917 		}
918 	}
919 	if (m == 0)
920 		change(file1, f1, file2, f2, 1, 0, 1, len[1], &flags);
921 	if (diff_format == D_IFDEF || diff_format == D_GFORMAT) {
922 		for (;;) {
923 #define	c i0
924 			if ((c = getc(f1)) == EOF)
925 				return;
926 			printf("%c", c);
927 		}
928 #undef c
929 	}
930 	if (anychange != 0) {
931 		if (diff_format == D_CONTEXT)
932 			dump_context_vec(f1, f2, flags);
933 		else if (diff_format == D_UNIFIED)
934 			dump_unified_vec(f1, f2, flags);
935 	}
936 }
937 
938 static void
939 range(int a, int b, const char *separator)
940 {
941 	printf("%d", a > b ? b : a);
942 	if (a < b)
943 		printf("%s%d", separator, b);
944 }
945 
946 static void
947 uni_range(int a, int b)
948 {
949 	if (a < b)
950 		printf("%d,%d", a, b - a + 1);
951 	else if (a == b)
952 		printf("%d", b);
953 	else
954 		printf("%d,0", b);
955 }
956 
957 static char *
958 preadline(int fd, size_t rlen, off_t off)
959 {
960 	char *line;
961 	ssize_t nr;
962 
963 	line = xmalloc(rlen + 1);
964 	if ((nr = pread(fd, line, rlen, off)) == -1)
965 		err(2, "preadline");
966 	if (nr > 0 && line[nr-1] == '\n')
967 		nr--;
968 	line[nr] = '\0';
969 	return (line);
970 }
971 
972 static bool
973 ignoreline_pattern(char *line)
974 {
975 	int ret;
976 
977 	ret = regexec(&ignore_re, line, 0, NULL, 0);
978 	return (ret == 0);	/* if it matched, it should be ignored. */
979 }
980 
981 static bool
982 ignoreline(char *line, bool skip_blanks)
983 {
984 
985 	if (skip_blanks && *line == '\0')
986 		return (true);
987 	if (ignore_pats != NULL && ignoreline_pattern(line))
988 		return (true);
989 	return (false);
990 }
991 
992 /*
993  * Indicate that there is a difference between lines a and b of the from file
994  * to get to lines c to d of the to file.  If a is greater then b then there
995  * are no lines in the from file involved and this means that there were
996  * lines appended (beginning at b).  If c is greater than d then there are
997  * lines missing from the to file.
998  */
999 static void
1000 change(char *file1, FILE *f1, char *file2, FILE *f2, int a, int b, int c, int d,
1001     int *pflags)
1002 {
1003 	static size_t max_context = 64;
1004 	long curpos;
1005 	int i, nc;
1006 	const char *walk;
1007 	bool skip_blanks, ignore;
1008 
1009 	skip_blanks = (*pflags & D_SKIPBLANKLINES);
1010 restart:
1011 	if ((diff_format != D_IFDEF || diff_format == D_GFORMAT) &&
1012 	    a > b && c > d)
1013 		return;
1014 	if (ignore_pats != NULL || skip_blanks) {
1015 		char *line;
1016 		/*
1017 		 * All lines in the change, insert, or delete must match an ignore
1018 		 * pattern for the change to be ignored.
1019 		 */
1020 		if (a <= b) {		/* Changes and deletes. */
1021 			for (i = a; i <= b; i++) {
1022 				line = preadline(fileno(f1),
1023 				    ixold[i] - ixold[i - 1], ixold[i - 1]);
1024 				ignore = ignoreline(line, skip_blanks);
1025 				free(line);
1026 				if (!ignore)
1027 					goto proceed;
1028 			}
1029 		}
1030 		if (a > b || c <= d) {	/* Changes and inserts. */
1031 			for (i = c; i <= d; i++) {
1032 				line = preadline(fileno(f2),
1033 				    ixnew[i] - ixnew[i - 1], ixnew[i - 1]);
1034 				ignore = ignoreline(line, skip_blanks);
1035 				free(line);
1036 				if (!ignore)
1037 					goto proceed;
1038 			}
1039 		}
1040 		return;
1041 	}
1042 proceed:
1043 	if (*pflags & D_HEADER && diff_format != D_BRIEF) {
1044 		printf("%s %s %s\n", diffargs, file1, file2);
1045 		*pflags &= ~D_HEADER;
1046 	}
1047 	if (diff_format == D_CONTEXT || diff_format == D_UNIFIED) {
1048 		/*
1049 		 * Allocate change records as needed.
1050 		 */
1051 		if (context_vec_start == NULL ||
1052 		    context_vec_ptr == context_vec_end - 1) {
1053 			ptrdiff_t offset = -1;
1054 
1055 			if (context_vec_start != NULL)
1056 				offset = context_vec_ptr - context_vec_start;
1057 			max_context <<= 1;
1058 			context_vec_start = xreallocarray(context_vec_start,
1059 			    max_context, sizeof(*context_vec_start));
1060 			context_vec_end = context_vec_start + max_context;
1061 			context_vec_ptr = context_vec_start + offset;
1062 		}
1063 		if (anychange == 0) {
1064 			/*
1065 			 * Print the context/unidiff header first time through.
1066 			 */
1067 			print_header(file1, file2);
1068 			anychange = 1;
1069 		} else if (a > context_vec_ptr->b + (2 * diff_context) + 1 &&
1070 		    c > context_vec_ptr->d + (2 * diff_context) + 1) {
1071 			/*
1072 			 * If this change is more than 'diff_context' lines from the
1073 			 * previous change, dump the record and reset it.
1074 			 */
1075 			if (diff_format == D_CONTEXT)
1076 				dump_context_vec(f1, f2, *pflags);
1077 			else
1078 				dump_unified_vec(f1, f2, *pflags);
1079 		}
1080 		context_vec_ptr++;
1081 		context_vec_ptr->a = a;
1082 		context_vec_ptr->b = b;
1083 		context_vec_ptr->c = c;
1084 		context_vec_ptr->d = d;
1085 		return;
1086 	}
1087 	if (anychange == 0)
1088 		anychange = 1;
1089 	switch (diff_format) {
1090 	case D_BRIEF:
1091 		return;
1092 	case D_NORMAL:
1093 	case D_EDIT:
1094 		range(a, b, ",");
1095 		printf("%c", a > b ? 'a' : c > d ? 'd' : 'c');
1096 		if (diff_format == D_NORMAL)
1097 			range(c, d, ",");
1098 		printf("\n");
1099 		break;
1100 	case D_REVERSE:
1101 		printf("%c", a > b ? 'a' : c > d ? 'd' : 'c');
1102 		range(a, b, " ");
1103 		printf("\n");
1104 		break;
1105 	case D_NREVERSE:
1106 		if (a > b)
1107 			printf("a%d %d\n", b, d - c + 1);
1108 		else {
1109 			printf("d%d %d\n", a, b - a + 1);
1110 			if (!(c > d))
1111 				/* add changed lines */
1112 				printf("a%d %d\n", b, d - c + 1);
1113 		}
1114 		break;
1115 	}
1116 	if (diff_format == D_GFORMAT) {
1117 		curpos = ftell(f1);
1118 		/* print through if append (a>b), else to (nb: 0 vs 1 orig) */
1119 		nc = ixold[a > b ? b : a - 1] - curpos;
1120 		for (i = 0; i < nc; i++)
1121 			printf("%c", getc(f1));
1122 		for (walk = group_format; *walk != '\0'; walk++) {
1123 			if (*walk == '%') {
1124 				walk++;
1125 				switch (*walk) {
1126 				case '<':
1127 					fetch(ixold, a, b, f1, '<', 1, *pflags);
1128 					break;
1129 				case '>':
1130 					fetch(ixnew, c, d, f2, '>', 0, *pflags);
1131 					break;
1132 				default:
1133 					printf("%%%c", *walk);
1134 					break;
1135 				}
1136 				continue;
1137 			}
1138 			printf("%c", *walk);
1139 		}
1140 	}
1141 	if (diff_format == D_SIDEBYSIDE) {
1142 		if (color && a > b)
1143 			printf("\033[%sm", add_code);
1144 		else if (color && c > d)
1145 			printf("\033[%sm", del_code);
1146 		if (a > b) {
1147 			print_space(0, hw + padding , *pflags);
1148 		} else {
1149 			nc = fetch(ixold, a, b, f1, '\0', 1, *pflags);
1150 			print_space(nc, hw - nc + padding, *pflags);
1151 		}
1152 		if (color && a > b)
1153 			printf("\033[%sm", add_code);
1154 		else if (color && c > d)
1155 			printf("\033[%sm", del_code);
1156 		printf("%c", (a > b) ? '>' : ((c > d) ? '<' : '|'));
1157 		if (color && c > d)
1158 			printf("\033[m");
1159 		print_space(hw + padding + 1 , padding, *pflags);
1160 		fetch(ixnew, c, d, f2, '\0', 0, *pflags);
1161 		printf("\n");
1162 	}
1163 	if (diff_format == D_NORMAL || diff_format == D_IFDEF) {
1164 		fetch(ixold, a, b, f1, '<', 1, *pflags);
1165 		if (a <= b && c <= d && diff_format == D_NORMAL)
1166 			printf("---\n");
1167 	}
1168 	if (diff_format != D_GFORMAT && diff_format != D_SIDEBYSIDE)
1169 		fetch(ixnew, c, d, f2, diff_format == D_NORMAL ? '>' : '\0', 0, *pflags);
1170 	if (edoffset != 0 && diff_format == D_EDIT) {
1171 		/*
1172 		 * A non-zero edoffset value for D_EDIT indicates that the last line
1173 		 * printed was a bare dot (".") that has been escaped as ".." to
1174 		 * prevent ed(1) from misinterpreting it.  We have to add a
1175 		 * substitute command to change this back and restart where we left
1176 		 * off.
1177 		 */
1178 		printf(".\n");
1179 		printf("%ds/.//\n", a + edoffset - 1);
1180 		b = a + edoffset - 1;
1181 		a = b + 1;
1182 		c += edoffset;
1183 		goto restart;
1184 	}
1185 	if ((diff_format == D_EDIT || diff_format == D_REVERSE) && c <= d)
1186 		printf(".\n");
1187 	if (inifdef) {
1188 		printf("#endif /* %s */\n", ifdefname);
1189 		inifdef = 0;
1190 	}
1191 }
1192 
1193 static int
1194 fetch(long *f, int a, int b, FILE *lb, int ch, int oldfile, int flags)
1195 {
1196 	int i, j, c, lastc, col, nc, newcol;
1197 
1198 	edoffset = 0;
1199 	nc = 0;
1200 	/*
1201 	 * When doing #ifdef's, copy down to current line
1202 	 * if this is the first file, so that stuff makes it to output.
1203 	 */
1204 	if ((diff_format == D_IFDEF) && oldfile) {
1205 		long curpos = ftell(lb);
1206 		/* print through if append (a>b), else to (nb: 0 vs 1 orig) */
1207 		nc = f[a > b ? b : a - 1] - curpos;
1208 		for (i = 0; i < nc; i++)
1209 			printf("%c", getc(lb));
1210 	}
1211 	if (a > b)
1212 		return (0);
1213 	if (diff_format == D_IFDEF) {
1214 		if (inifdef) {
1215 			printf("#else /* %s%s */\n",
1216 			    oldfile == 1 ? "!" : "", ifdefname);
1217 		} else {
1218 			if (oldfile)
1219 				printf("#ifndef %s\n", ifdefname);
1220 			else
1221 				printf("#ifdef %s\n", ifdefname);
1222 		}
1223 		inifdef = 1 + oldfile;
1224 	}
1225 	for (i = a; i <= b; i++) {
1226 		fseek(lb, f[i - 1], SEEK_SET);
1227 		nc = f[i] - f[i - 1];
1228 		if (diff_format == D_SIDEBYSIDE && hw < nc)
1229 			nc = hw;
1230 		if (diff_format != D_IFDEF && diff_format != D_GFORMAT &&
1231 		    ch != '\0') {
1232 			if (color && (ch == '>' || ch == '+'))
1233 				printf("\033[%sm", add_code);
1234 			else if (color && (ch == '<' || ch == '-'))
1235 				printf("\033[%sm", del_code);
1236 			printf("%c", ch);
1237 			if (Tflag && (diff_format == D_NORMAL ||
1238 			    diff_format == D_CONTEXT ||
1239 			    diff_format == D_UNIFIED))
1240 				printf("\t");
1241 			else if (diff_format != D_UNIFIED)
1242 				printf(" ");
1243 		}
1244 		col = j = 0;
1245 		lastc = '\0';
1246 		while (j < nc && (hw == 0 || col < hw)) {
1247 			c = getc(lb);
1248 			if (flags & D_STRIPCR && c == '\r') {
1249 				if ((c = getc(lb)) == '\n')
1250 					j++;
1251 				else {
1252 					ungetc(c, lb);
1253 					c = '\r';
1254 				}
1255 			}
1256 			if (c == EOF) {
1257 				if (diff_format == D_EDIT ||
1258 				    diff_format == D_REVERSE ||
1259 				    diff_format == D_NREVERSE)
1260 					warnx("No newline at end of file");
1261 				else
1262 					printf("\n\\ No newline at end of file\n");
1263 				return (col);
1264 			}
1265 			/*
1266 			 * when using --side-by-side, col needs to be increased
1267 			 * in any case to keep the columns aligned
1268 			 */
1269 			if (c == '\t') {
1270 				if (flags & D_EXPANDTABS) {
1271 					newcol = ((col / tabsize) + 1) * tabsize;
1272 					do {
1273 						printf(" ");
1274 					} while (++col < newcol && col < hw);
1275 				} else {
1276 					if (diff_format == D_SIDEBYSIDE) {
1277 						if ((col + tabsize) > hw) {
1278 							printf("%*s", hw - col, "");
1279 							col = hw;
1280 						} else {
1281 							printf("\t");
1282 							col += tabsize - 1;
1283 						}
1284 					} else {
1285 						printf("\t");
1286 						col++;
1287 					}
1288 				}
1289 			} else {
1290 				if (diff_format == D_EDIT && j == 1 && c == '\n' &&
1291 				    lastc == '.') {
1292 					/*
1293 					 * Don't print a bare "." line since that will confuse
1294 					 * ed(1). Print ".." instead and set the, global variable
1295 					 * edoffset to an offset from which to restart. The
1296 					 * caller must check the value of edoffset
1297 					 */
1298 					printf(".\n");
1299 					edoffset = i - a + 1;
1300 					return (edoffset);
1301 				}
1302 				/* when side-by-side, do not print a newline */
1303 				if (diff_format != D_SIDEBYSIDE || c != '\n') {
1304 					if (color && c == '\n')
1305 						printf("\033[m%c", c);
1306 					else
1307 						printf("%c", c);
1308 					col++;
1309 				}
1310 			}
1311 
1312 			j++;
1313 			lastc = c;
1314 		}
1315 	}
1316 	if (color && diff_format == D_SIDEBYSIDE)
1317 		printf("\033[m");
1318 	return (col);
1319 }
1320 
1321 /*
1322  * Hash function taken from Robert Sedgewick, Algorithms in C, 3d ed., p 578.
1323  */
1324 static enum readhash
1325 readhash(FILE *f, int flags, unsigned *hash)
1326 {
1327 	int i, t, space;
1328 	unsigned sum;
1329 
1330 	sum = 1;
1331 	space = 0;
1332 	for (i = 0;;) {
1333 		switch (t = getc(f)) {
1334 		case '\0':
1335 			if ((flags & D_FORCEASCII) == 0)
1336 				return (RH_BINARY);
1337 			goto hashchar;
1338 		case '\r':
1339 			if (flags & D_STRIPCR) {
1340 				t = getc(f);
1341 				if (t == '\n')
1342 					break;
1343 				ungetc(t, f);
1344 			}
1345 			/* FALLTHROUGH */
1346 		case '\t':
1347 		case '\v':
1348 		case '\f':
1349 		case ' ':
1350 			if ((flags & (D_FOLDBLANKS|D_IGNOREBLANKS)) != 0) {
1351 				space++;
1352 				continue;
1353 			}
1354 			/* FALLTHROUGH */
1355 		default:
1356 		hashchar:
1357 			if (space && (flags & D_IGNOREBLANKS) == 0) {
1358 				i++;
1359 				space = 0;
1360 			}
1361 			sum = sum * 127 + chrtran(t);
1362 			i++;
1363 			continue;
1364 		case EOF:
1365 			if (i == 0)
1366 				return (RH_EOF);
1367 			/* FALLTHROUGH */
1368 		case '\n':
1369 			break;
1370 		}
1371 		break;
1372 	}
1373 	*hash = sum;
1374 	return (RH_OK);
1375 }
1376 
1377 static int
1378 asciifile(FILE *f)
1379 {
1380 	unsigned char buf[BUFSIZ];
1381 	size_t cnt;
1382 
1383 	if (f == NULL)
1384 		return (1);
1385 
1386 	rewind(f);
1387 	cnt = fread(buf, 1, sizeof(buf), f);
1388 	return (memchr(buf, '\0', cnt) == NULL);
1389 }
1390 
1391 #define begins_with(s, pre) (strncmp(s, pre, sizeof(pre) - 1) == 0)
1392 
1393 static char *
1394 match_function(const long *f, int pos, FILE *fp)
1395 {
1396 	unsigned char buf[FUNCTION_CONTEXT_SIZE];
1397 	size_t nc;
1398 	int last = lastline;
1399 	const char *state = NULL;
1400 
1401 	lastline = pos;
1402 	for (; pos > last; pos--) {
1403 		fseek(fp, f[pos - 1], SEEK_SET);
1404 		nc = f[pos] - f[pos - 1];
1405 		if (nc >= sizeof(buf))
1406 			nc = sizeof(buf) - 1;
1407 		nc = fread(buf, 1, nc, fp);
1408 		if (nc == 0)
1409 			continue;
1410 		buf[nc] = '\0';
1411 		buf[strcspn(buf, "\n")] = '\0';
1412 		if (most_recent_pat != NULL) {
1413 			int ret = regexec(&most_recent_re, buf, 0, NULL, 0);
1414 
1415 			if (ret != 0)
1416 				continue;
1417 			strlcpy(lastbuf, buf, sizeof(lastbuf));
1418 			lastmatchline = pos;
1419 			return (lastbuf);
1420 		} else if (isalpha(buf[0]) || buf[0] == '_' || buf[0] == '$'
1421 			|| buf[0] == '-' || buf[0] == '+') {
1422 			if (begins_with(buf, "private:")) {
1423 				if (!state)
1424 					state = " (private)";
1425 			} else if (begins_with(buf, "protected:")) {
1426 				if (!state)
1427 					state = " (protected)";
1428 			} else if (begins_with(buf, "public:")) {
1429 				if (!state)
1430 					state = " (public)";
1431 			} else {
1432 				strlcpy(lastbuf, buf, sizeof(lastbuf));
1433 				if (state)
1434 					strlcat(lastbuf, state, sizeof(lastbuf));
1435 				lastmatchline = pos;
1436 				return (lastbuf);
1437 			}
1438 		}
1439 	}
1440 	return (lastmatchline > 0 ? lastbuf : NULL);
1441 }
1442 
1443 /* dump accumulated "context" diff changes */
1444 static void
1445 dump_context_vec(FILE *f1, FILE *f2, int flags)
1446 {
1447 	struct context_vec *cvp = context_vec_start;
1448 	int lowa, upb, lowc, upd, do_output;
1449 	int a, b, c, d;
1450 	char ch, *f;
1451 
1452 	if (context_vec_start > context_vec_ptr)
1453 		return;
1454 
1455 	b = d = 0;		/* gcc */
1456 	lowa = MAX(1, cvp->a - diff_context);
1457 	upb = MIN(len[0], context_vec_ptr->b + diff_context);
1458 	lowc = MAX(1, cvp->c - diff_context);
1459 	upd = MIN(len[1], context_vec_ptr->d + diff_context);
1460 
1461 	printf("***************");
1462 	if (flags & (D_PROTOTYPE | D_MATCHLAST)) {
1463 		f = match_function(ixold, cvp->a - 1, f1);
1464 		if (f != NULL)
1465 			printf(" %s", f);
1466 	}
1467 	printf("\n*** ");
1468 	range(lowa, upb, ",");
1469 	printf(" ****\n");
1470 
1471 	/*
1472 	 * Output changes to the "old" file.  The first loop suppresses
1473 	 * output if there were no changes to the "old" file (we'll see
1474 	 * the "old" lines as context in the "new" list).
1475 	 */
1476 	do_output = 0;
1477 	for (; cvp <= context_vec_ptr; cvp++)
1478 		if (cvp->a <= cvp->b) {
1479 			cvp = context_vec_start;
1480 			do_output++;
1481 			break;
1482 		}
1483 	if (do_output) {
1484 		while (cvp <= context_vec_ptr) {
1485 			a = cvp->a;
1486 			b = cvp->b;
1487 			c = cvp->c;
1488 			d = cvp->d;
1489 
1490 			if (a <= b && c <= d)
1491 				ch = 'c';
1492 			else
1493 				ch = (a <= b) ? 'd' : 'a';
1494 
1495 			if (ch == 'a')
1496 				fetch(ixold, lowa, b, f1, ' ', 0, flags);
1497 			else {
1498 				fetch(ixold, lowa, a - 1, f1, ' ', 0, flags);
1499 				fetch(ixold, a, b, f1,
1500 				    ch == 'c' ? '!' : '-', 0, flags);
1501 			}
1502 			lowa = b + 1;
1503 			cvp++;
1504 		}
1505 		fetch(ixold, b + 1, upb, f1, ' ', 0, flags);
1506 	}
1507 	/* output changes to the "new" file */
1508 	printf("--- ");
1509 	range(lowc, upd, ",");
1510 	printf(" ----\n");
1511 
1512 	do_output = 0;
1513 	for (cvp = context_vec_start; cvp <= context_vec_ptr; cvp++)
1514 		if (cvp->c <= cvp->d) {
1515 			cvp = context_vec_start;
1516 			do_output++;
1517 			break;
1518 		}
1519 	if (do_output) {
1520 		while (cvp <= context_vec_ptr) {
1521 			a = cvp->a;
1522 			b = cvp->b;
1523 			c = cvp->c;
1524 			d = cvp->d;
1525 
1526 			if (a <= b && c <= d)
1527 				ch = 'c';
1528 			else
1529 				ch = (a <= b) ? 'd' : 'a';
1530 
1531 			if (ch == 'd')
1532 				fetch(ixnew, lowc, d, f2, ' ', 0, flags);
1533 			else {
1534 				fetch(ixnew, lowc, c - 1, f2, ' ', 0, flags);
1535 				fetch(ixnew, c, d, f2,
1536 				    ch == 'c' ? '!' : '+', 0, flags);
1537 			}
1538 			lowc = d + 1;
1539 			cvp++;
1540 		}
1541 		fetch(ixnew, d + 1, upd, f2, ' ', 0, flags);
1542 	}
1543 	context_vec_ptr = context_vec_start - 1;
1544 }
1545 
1546 /* dump accumulated "unified" diff changes */
1547 static void
1548 dump_unified_vec(FILE *f1, FILE *f2, int flags)
1549 {
1550 	struct context_vec *cvp = context_vec_start;
1551 	int lowa, upb, lowc, upd;
1552 	int a, b, c, d;
1553 	char ch, *f;
1554 
1555 	if (context_vec_start > context_vec_ptr)
1556 		return;
1557 
1558 	b = d = 0;		/* gcc */
1559 	lowa = MAX(1, cvp->a - diff_context);
1560 	upb = MIN(len[0], context_vec_ptr->b + diff_context);
1561 	lowc = MAX(1, cvp->c - diff_context);
1562 	upd = MIN(len[1], context_vec_ptr->d + diff_context);
1563 
1564 	printf("@@ -");
1565 	uni_range(lowa, upb);
1566 	printf(" +");
1567 	uni_range(lowc, upd);
1568 	printf(" @@");
1569 	if (flags & (D_PROTOTYPE | D_MATCHLAST)) {
1570 		f = match_function(ixold, cvp->a - 1, f1);
1571 		if (f != NULL)
1572 			printf(" %s", f);
1573 	}
1574 	printf("\n");
1575 
1576 	/*
1577 	 * Output changes in "unified" diff format--the old and new lines
1578 	 * are printed together.
1579 	 */
1580 	for (; cvp <= context_vec_ptr; cvp++) {
1581 		a = cvp->a;
1582 		b = cvp->b;
1583 		c = cvp->c;
1584 		d = cvp->d;
1585 
1586 		/*
1587 		 * c: both new and old changes
1588 		 * d: only changes in the old file
1589 		 * a: only changes in the new file
1590 		 */
1591 		if (a <= b && c <= d)
1592 			ch = 'c';
1593 		else
1594 			ch = (a <= b) ? 'd' : 'a';
1595 
1596 		switch (ch) {
1597 		case 'c':
1598 			fetch(ixold, lowa, a - 1, f1, ' ', 0, flags);
1599 			fetch(ixold, a, b, f1, '-', 0, flags);
1600 			fetch(ixnew, c, d, f2, '+', 0, flags);
1601 			break;
1602 		case 'd':
1603 			fetch(ixold, lowa, a - 1, f1, ' ', 0, flags);
1604 			fetch(ixold, a, b, f1, '-', 0, flags);
1605 			break;
1606 		case 'a':
1607 			fetch(ixnew, lowc, c - 1, f2, ' ', 0, flags);
1608 			fetch(ixnew, c, d, f2, '+', 0, flags);
1609 			break;
1610 		}
1611 		lowa = b + 1;
1612 		lowc = d + 1;
1613 	}
1614 	fetch(ixnew, d + 1, upd, f2, ' ', 0, flags);
1615 
1616 	context_vec_ptr = context_vec_start - 1;
1617 }
1618 
1619 static void
1620 print_header(const char *file1, const char *file2)
1621 {
1622 	const char *time_format;
1623 	char buf[256];
1624 	struct tm tm1, tm2, *tm_ptr1, *tm_ptr2;
1625 	int nsec1 = stb1.st_mtim.tv_nsec;
1626 	int nsec2 = stb2.st_mtim.tv_nsec;
1627 
1628 	time_format = "%Y-%m-%d %H:%M:%S";
1629 
1630 	if (cflag)
1631 		time_format = "%c";
1632 	tm_ptr1 = localtime_r(&stb1.st_mtime, &tm1);
1633 	tm_ptr2 = localtime_r(&stb2.st_mtime, &tm2);
1634 	if (label[0] != NULL)
1635 		printf("%s %s\n", diff_format == D_CONTEXT ? "***" : "---",
1636 		    label[0]);
1637 	else {
1638 		strftime(buf, sizeof(buf), time_format, tm_ptr1);
1639 		printf("%s %s\t%s", diff_format == D_CONTEXT ? "***" : "---",
1640 		    file1, buf);
1641 		if (!cflag) {
1642 			strftime(buf, sizeof(buf), "%z", tm_ptr1);
1643 			printf(".%.9d %s", nsec1, buf);
1644 		}
1645 		printf("\n");
1646 	}
1647 	if (label[1] != NULL)
1648 		printf("%s %s\n", diff_format == D_CONTEXT ? "---" : "+++",
1649 		    label[1]);
1650 	else {
1651 		strftime(buf, sizeof(buf), time_format, tm_ptr2);
1652 		printf("%s %s\t%s", diff_format == D_CONTEXT ? "---" : "+++",
1653 		    file2, buf);
1654 		if (!cflag) {
1655 			strftime(buf, sizeof(buf), "%z", tm_ptr2);
1656 			printf(".%.9d %s", nsec2, buf);
1657 		}
1658 		printf("\n");
1659 	}
1660 }
1661 
1662 /*
1663  * Prints n number of space characters either by using tab
1664  * or single space characters.
1665  * nc is the preceding number of characters
1666  */
1667 static void
1668 print_space(int nc, int n, int flags) {
1669 	int i, col;
1670 
1671 	col = n;
1672 	if ((flags & D_EXPANDTABS) == 0) {
1673 		/* first tabstop may be closer than tabsize */
1674 		i = tabsize - (nc % tabsize);
1675 		while (col >= tabsize) {
1676 			printf("\t");
1677 			col -= i;
1678 			i = tabsize;
1679 		}
1680 	}
1681 	printf("%*s", col, "");
1682 }
1683