1/*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1991, 1993
5 *	The Regents of the University of California.  All rights reserved.
6 *
7 * This code is derived from software contributed to Berkeley by
8 * David Hitz of Auspex Systems, Inc.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 *    notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 *    notice, this list of conditions and the following disclaimer in the
17 *    documentation and/or other materials provided with the distribution.
18 * 3. Neither the name of the University nor the names of its contributors
19 *    may be used to endorse or promote products derived from this software
20 *    without specific prior written permission.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * SUCH DAMAGE.
33 */
34
35#ifndef lint
36static const char copyright[] =
37"@(#) Copyright (c) 1991, 1993\n\
38	The Regents of the University of California.  All rights reserved.\n";
39#endif /* not lint */
40
41#ifndef lint
42#if 0
43static char sccsid[] = "@(#)look.c	8.2 (Berkeley) 5/4/95";
44#endif
45#endif /* not lint */
46#include <sys/cdefs.h>
47__FBSDID("$FreeBSD$");
48
49/*
50 * look -- find lines in a sorted list.
51 *
52 * The man page said that TABs and SPACEs participate in -d comparisons.
53 * In fact, they were ignored.  This implements historic practice, not
54 * the manual page.
55 */
56
57#include <sys/types.h>
58#include <sys/mman.h>
59#include <sys/stat.h>
60
61#include <err.h>
62#include <errno.h>
63#include <fcntl.h>
64#include <getopt.h>
65#include <limits.h>
66#include <locale.h>
67#include <stdint.h>
68#include <stdio.h>
69#include <stdlib.h>
70#include <string.h>
71#include <unistd.h>
72#include <wchar.h>
73#include <wctype.h>
74
75#include "pathnames.h"
76
77static char _path_words[] = _PATH_WORDS;
78
79#define	EQUAL		0
80#define	GREATER		1
81#define	LESS		(-1)
82
83static int dflag, fflag;
84
85static char	*binary_search(wchar_t *, unsigned char *, unsigned char *);
86static int	 compare(wchar_t *, unsigned char *, unsigned char *);
87static char	*linear_search(wchar_t *, unsigned char *, unsigned char *);
88static int	 look(wchar_t *, unsigned char *, unsigned char *);
89static wchar_t	*prepkey(const char *, wchar_t);
90static void	 print_from(wchar_t *, unsigned char *, unsigned char *);
91
92static void usage(void);
93
94static struct option longopts[] = {
95	{ "alternative",no_argument,	NULL, 'a' },
96	{ "alphanum",	no_argument,	NULL, 'd' },
97	{ "ignore-case",no_argument,	NULL, 'i' },
98	{ "terminate",	required_argument, NULL, 't'},
99	{ NULL,		0,		NULL, 0 },
100};
101
102int
103main(int argc, char *argv[])
104{
105	struct stat sb;
106	int ch, fd, match;
107	wchar_t termchar;
108	unsigned char *back, *front;
109	unsigned const char *file;
110	wchar_t *key;
111
112	(void) setlocale(LC_CTYPE, "");
113
114	file = _path_words;
115	termchar = L'\0';
116	while ((ch = getopt_long(argc, argv, "+adft:", longopts, NULL)) != -1)
117		switch(ch) {
118		case 'a':
119			/* COMPATIBILITY */
120			break;
121		case 'd':
122			dflag = 1;
123			break;
124		case 'f':
125			fflag = 1;
126			break;
127		case 't':
128			if (mbrtowc(&termchar, optarg, MB_LEN_MAX, NULL) !=
129			    strlen(optarg))
130				errx(2, "invalid termination character");
131			break;
132		case '?':
133		default:
134			usage();
135		}
136	argc -= optind;
137	argv += optind;
138
139	if (argc == 0)
140		usage();
141	if (argc == 1) 			/* But set -df by default. */
142		dflag = fflag = 1;
143	key = prepkey(*argv++, termchar);
144	if (argc >= 2)
145		file = *argv++;
146
147	match = 1;
148
149	do {
150		if ((fd = open(file, O_RDONLY, 0)) < 0 || fstat(fd, &sb))
151			err(2, "%s", file);
152		if ((uintmax_t)sb.st_size > (uintmax_t)SIZE_T_MAX)
153			errx(2, "%s: %s", file, strerror(EFBIG));
154		if (sb.st_size == 0) {
155			close(fd);
156			continue;
157		}
158		if ((front = mmap(NULL, (size_t)sb.st_size, PROT_READ, MAP_SHARED, fd, (off_t)0)) == MAP_FAILED)
159			err(2, "%s", file);
160		back = front + sb.st_size;
161		match *= (look(key, front, back));
162		close(fd);
163	} while (argc-- > 2 && (file = *argv++));
164
165	exit(match);
166}
167
168static wchar_t *
169prepkey(const char *string, wchar_t termchar)
170{
171	const char *readp;
172	wchar_t *key, *writep;
173	wchar_t ch;
174	size_t clen;
175
176	/*
177	 * Reformat search string and convert to wide character representation
178	 * to avoid doing it multiple times later.
179	 */
180	if ((key = malloc(sizeof(wchar_t) * (strlen(string) + 1))) == NULL)
181		err(2, NULL);
182	readp = string;
183	writep = key;
184	while ((clen = mbrtowc(&ch, readp, MB_LEN_MAX, NULL)) != 0) {
185		if (clen == (size_t)-1 || clen == (size_t)-2)
186			errc(2, EILSEQ, NULL);
187		if (fflag)
188			ch = towlower(ch);
189		if (!dflag || iswalnum(ch))
190			*writep++ = ch;
191		readp += clen;
192	}
193	*writep = L'\0';
194	if (termchar != L'\0' && (writep = wcschr(key, termchar)) != NULL)
195		*++writep = L'\0';
196	return (key);
197}
198
199static int
200look(wchar_t *string, unsigned char *front, unsigned char *back)
201{
202
203	front = binary_search(string, front, back);
204	front = linear_search(string, front, back);
205
206	if (front)
207		print_from(string, front, back);
208	return (front ? 0 : 1);
209}
210
211
212/*
213 * Binary search for "string" in memory between "front" and "back".
214 *
215 * This routine is expected to return a pointer to the start of a line at
216 * *or before* the first word matching "string".  Relaxing the constraint
217 * this way simplifies the algorithm.
218 *
219 * Invariants:
220 * 	front points to the beginning of a line at or before the first
221 *	matching string.
222 *
223 * 	back points to the beginning of a line at or after the first
224 *	matching line.
225 *
226 * Base of the Invariants.
227 * 	front = NULL;
228 *	back = EOF;
229 *
230 * Advancing the Invariants:
231 *
232 * 	p = first newline after halfway point from front to back.
233 *
234 * 	If the string at "p" is not greater than the string to match,
235 *	p is the new front.  Otherwise it is the new back.
236 *
237 * Termination:
238 *
239 * 	The definition of the routine allows it return at any point,
240 *	since front is always at or before the line to print.
241 *
242 * 	In fact, it returns when the chosen "p" equals "back".  This
243 *	implies that there exists a string is least half as long as
244 *	(back - front), which in turn implies that a linear search will
245 *	be no more expensive than the cost of simply printing a string or two.
246 *
247 * 	Trying to continue with binary search at this point would be
248 *	more trouble than it's worth.
249 */
250#define	SKIP_PAST_NEWLINE(p, back) \
251	while (p < back && *p++ != '\n');
252
253static char *
254binary_search(wchar_t *string, unsigned char *front, unsigned char *back)
255{
256	unsigned char *p;
257
258	p = front + (back - front) / 2;
259	SKIP_PAST_NEWLINE(p, back);
260
261	/*
262	 * If the file changes underneath us, make sure we don't
263	 * infinitely loop.
264	 */
265	while (p < back && back > front) {
266		if (compare(string, p, back) == GREATER)
267			front = p;
268		else
269			back = p;
270		p = front + (back - front) / 2;
271		SKIP_PAST_NEWLINE(p, back);
272	}
273	return (front);
274}
275
276/*
277 * Find the first line that starts with string, linearly searching from front
278 * to back.
279 *
280 * Return NULL for no such line.
281 *
282 * This routine assumes:
283 *
284 * 	o front points at the first character in a line.
285 *	o front is before or at the first line to be printed.
286 */
287static char *
288linear_search(wchar_t *string, unsigned char *front, unsigned char *back)
289{
290	while (front < back) {
291		switch (compare(string, front, back)) {
292		case EQUAL:		/* Found it. */
293			return (front);
294		case LESS:		/* No such string. */
295			return (NULL);
296		case GREATER:		/* Keep going. */
297			break;
298		}
299		SKIP_PAST_NEWLINE(front, back);
300	}
301	return (NULL);
302}
303
304/*
305 * Print as many lines as match string, starting at front.
306 */
307static void
308print_from(wchar_t *string, unsigned char *front, unsigned char *back)
309{
310	for (; front < back && compare(string, front, back) == EQUAL; ++front) {
311		for (; front < back && *front != '\n'; ++front)
312			if (putchar(*front) == EOF)
313				err(2, "stdout");
314		if (putchar('\n') == EOF)
315			err(2, "stdout");
316	}
317}
318
319/*
320 * Return LESS, GREATER, or EQUAL depending on how the string1 compares with
321 * string2 (s1 ??? s2).
322 *
323 * 	o Matches up to len(s1) are EQUAL.
324 *	o Matches up to len(s2) are GREATER.
325 *
326 * Compare understands about the -f and -d flags, and treats comparisons
327 * appropriately.
328 *
329 * The string "s1" is null terminated.  The string s2 is '\n' terminated (or
330 * "back" terminated).
331 */
332static int
333compare(wchar_t *s1, unsigned char *s2, unsigned char *back)
334{
335	wchar_t ch1, ch2;
336	size_t len2;
337
338	for (; *s1 && s2 < back && *s2 != '\n'; ++s1, s2 += len2) {
339		ch1 = *s1;
340		len2 = mbrtowc(&ch2, s2, back - s2, NULL);
341		if (len2 == (size_t)-1 || len2 == (size_t)-2) {
342			ch2 = *s2;
343			len2 = 1;
344		}
345		if (fflag)
346			ch2 = towlower(ch2);
347		if (dflag && !iswalnum(ch2)) {
348			/* Ignore character in comparison. */
349			--s1;
350			continue;
351		}
352		if (ch1 != ch2)
353			return (ch1 < ch2 ? LESS : GREATER);
354	}
355	return (*s1 ? GREATER : EQUAL);
356}
357
358static void
359usage(void)
360{
361	(void)fprintf(stderr, "usage: look [-df] [-t char] string [file ...]\n");
362	exit(2);
363}
364