1 /*	$OpenBSD: util.c,v 1.36 2007/10/02 17:59:18 otto Exp $	*/
2 /*	$FreeBSD: head/usr.bin/grep/fastgrep.c 211496 2010-08-19 09:28:59Z des $ */
3 
4 /*-
5  * Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav
6  * Copyright (C) 2008 Gabor Kovesdan <gabor@FreeBSD.org>
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 must retain the above copyright
13  *    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  *
18  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
22  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28  * SUCH DAMAGE.
29  */
30 
31 /*
32  * XXX: This file is a speed up for grep to cover the defects of the
33  * regex library.  These optimizations should practically be implemented
34  * there keeping this code clean.  This is a future TODO, but for the
35  * meantime, we need to use this workaround.
36  */
37 
38 #if HAVE_NBTOOL_CONFIG_H
39 #include "nbtool_config.h"
40 #endif
41 
42 #include <sys/cdefs.h>
43 __RCSID("$NetBSD: fastgrep.c,v 1.5 2011/04/18 03:27:40 joerg Exp $");
44 
45 #include <limits.h>
46 #include <stdbool.h>
47 #include <stdlib.h>
48 #include <string.h>
49 #include <wchar.h>
50 #include <wctype.h>
51 
52 #include "grep.h"
53 
54 static inline int	grep_cmp(const unsigned char *, const unsigned char *, size_t);
55 static inline void	grep_revstr(unsigned char *, int);
56 
57 void
fgrepcomp(fastgrep_t * fg,const char * pat)58 fgrepcomp(fastgrep_t *fg, const char *pat)
59 {
60 	unsigned int i;
61 
62 	/* Initialize. */
63 	fg->len = strlen(pat);
64 	fg->bol = false;
65 	fg->eol = false;
66 	fg->reversed = false;
67 
68 	fg->pattern = (unsigned char *)grep_strdup(pat);
69 
70 	/* Preprocess pattern. */
71 	for (i = 0; i <= UCHAR_MAX; i++)
72 		fg->qsBc[i] = fg->len;
73 	for (i = 1; i < fg->len; i++)
74 		fg->qsBc[fg->pattern[i]] = fg->len - i;
75 }
76 
77 /*
78  * Returns: -1 on failure, 0 on success
79  */
80 int
fastcomp(fastgrep_t * fg,const char * pat)81 fastcomp(fastgrep_t *fg, const char *pat)
82 {
83 	unsigned int i;
84 	int firstHalfDot = -1;
85 	int firstLastHalfDot = -1;
86 	int hasDot = 0;
87 	int lastHalfDot = 0;
88 	int shiftPatternLen;
89 
90 	/* Initialize. */
91 	fg->len = strlen(pat);
92 	fg->bol = false;
93 	fg->eol = false;
94 	fg->reversed = false;
95 	fg->word = wflag;
96 
97 	/* Remove end-of-line character ('$'). */
98 	if (fg->len > 0 && pat[fg->len - 1] == '$') {
99 		fg->eol = true;
100 		fg->len--;
101 	}
102 
103 	/* Remove beginning-of-line character ('^'). */
104 	if (pat[0] == '^') {
105 		fg->bol = true;
106 		fg->len--;
107 		pat++;
108 	}
109 
110 	if (fg->len >= 14 &&
111 	    memcmp(pat, "[[:<:]]", 7) == 0 &&
112 	    memcmp(pat + fg->len - 7, "[[:>:]]", 7) == 0) {
113 		fg->len -= 14;
114 		pat += 7;
115 		/* Word boundary is handled separately in util.c */
116 		fg->word = true;
117 	}
118 
119 	/*
120 	 * pat has been adjusted earlier to not include '^', '$' or
121 	 * the word match character classes at the beginning and ending
122 	 * of the string respectively.
123 	 */
124 	fg->pattern = grep_malloc(fg->len + 1);
125 	memcpy(fg->pattern, pat, fg->len);
126 	fg->pattern[fg->len] = '\0';
127 
128 	/* Look for ways to cheat...er...avoid the full regex engine. */
129 	for (i = 0; i < fg->len; i++) {
130 		/* Can still cheat? */
131 		if (fg->pattern[i] == '.') {
132 			hasDot = i;
133 			if (i < fg->len / 2) {
134 				if (firstHalfDot < 0)
135 					/* Closest dot to the beginning */
136 					firstHalfDot = i;
137 			} else {
138 				/* Closest dot to the end of the pattern. */
139 				lastHalfDot = i;
140 				if (firstLastHalfDot < 0)
141 					firstLastHalfDot = i;
142 			}
143 		} else {
144 			/* Free memory and let others know this is empty. */
145 			free(fg->pattern);
146 			fg->pattern = NULL;
147 			return (-1);
148 		}
149 	}
150 
151 	/*
152 	 * Determine if a reverse search would be faster based on the placement
153 	 * of the dots.
154 	 */
155 	if ((!(lflag || cflag)) && ((!(fg->bol || fg->eol)) &&
156 	    ((lastHalfDot) && ((firstHalfDot < 0) ||
157 	    ((fg->len - (lastHalfDot + 1)) < (size_t)firstHalfDot)))) &&
158 	    !oflag && !color) {
159 		fg->reversed = true;
160 		hasDot = fg->len - (firstHalfDot < 0 ?
161 		    firstLastHalfDot : firstHalfDot) - 1;
162 		grep_revstr(fg->pattern, fg->len);
163 	}
164 
165 	/*
166 	 * Normal Quick Search would require a shift based on the position the
167 	 * next character after the comparison is within the pattern.  With
168 	 * wildcards, the position of the last dot effects the maximum shift
169 	 * distance.
170 	 * The closer to the end the wild card is the slower the search.  A
171 	 * reverse version of this algorithm would be useful for wildcards near
172 	 * the end of the string.
173 	 *
174 	 * Examples:
175 	 * Pattern	Max shift
176 	 * -------	---------
177 	 * this		5
178 	 * .his		4
179 	 * t.is		3
180 	 * th.s		2
181 	 * thi.		1
182 	 */
183 
184 	/* Adjust the shift based on location of the last dot ('.'). */
185 	shiftPatternLen = fg->len - hasDot;
186 
187 	/* Preprocess pattern. */
188 	for (i = 0; i <= (signed)UCHAR_MAX; i++)
189 		fg->qsBc[i] = shiftPatternLen;
190 	for (i = hasDot + 1; i < fg->len; i++) {
191 		fg->qsBc[fg->pattern[i]] = fg->len - i;
192 	}
193 
194 	/*
195 	 * Put pattern back to normal after pre-processing to allow for easy
196 	 * comparisons later.
197 	 */
198 	if (fg->reversed)
199 		grep_revstr(fg->pattern, fg->len);
200 
201 	return (0);
202 }
203 
204 int
grep_search(fastgrep_t * fg,const unsigned char * data,size_t len,regmatch_t * pmatch)205 grep_search(fastgrep_t *fg, const unsigned char *data, size_t len, regmatch_t *pmatch)
206 {
207 	unsigned int j;
208 	int ret = REG_NOMATCH;
209 
210 	if (pmatch->rm_so == (ssize_t)len)
211 		return (ret);
212 
213 	if (fg->bol && pmatch->rm_so != 0) {
214 		pmatch->rm_so = len;
215 		pmatch->rm_eo = len;
216 		return (ret);
217 	}
218 
219 	/* No point in going farther if we do not have enough data. */
220 	if (len < fg->len)
221 		return (ret);
222 
223 	/* Only try once at the beginning or ending of the line. */
224 	if (fg->bol || fg->eol) {
225 		/* Simple text comparison. */
226 		/* Verify data is >= pattern length before searching on it. */
227 		if (len >= fg->len) {
228 			/* Determine where in data to start search at. */
229 			j = fg->eol ? len - fg->len : 0;
230 			if (!((fg->bol && fg->eol) && (len != fg->len)))
231 				if (grep_cmp(fg->pattern, data + j,
232 				    fg->len) == -1) {
233 					pmatch->rm_so = j;
234 					pmatch->rm_eo = j + fg->len;
235 						ret = 0;
236 				}
237 		}
238 	} else if (fg->reversed) {
239 		/* Quick Search algorithm. */
240 		j = len;
241 		do {
242 			if (grep_cmp(fg->pattern, data + j - fg->len,
243 				fg->len) == -1) {
244 				pmatch->rm_so = j - fg->len;
245 				pmatch->rm_eo = j;
246 				ret = 0;
247 				break;
248 			}
249 			/* Shift if within bounds, otherwise, we are done. */
250 			if (j == fg->len)
251 				break;
252 			j -= fg->qsBc[data[j - fg->len - 1]];
253 		} while (j >= fg->len);
254 	} else {
255 		/* Quick Search algorithm. */
256 		j = pmatch->rm_so;
257 		do {
258 			if (grep_cmp(fg->pattern, data + j, fg->len) == -1) {
259 				pmatch->rm_so = j;
260 				pmatch->rm_eo = j + fg->len;
261 				ret = 0;
262 				break;
263 			}
264 
265 			/* Shift if within bounds, otherwise, we are done. */
266 			if (j + fg->len == len)
267 				break;
268 			else
269 				j += fg->qsBc[data[j + fg->len]];
270 		} while (j <= (len - fg->len));
271 	}
272 
273 	return (ret);
274 }
275 
276 /*
277  * Returns:	i >= 0 on failure (position that it failed)
278  *		-1 on success
279  */
280 static inline int
grep_cmp(const unsigned char * pat,const unsigned char * data,size_t len)281 grep_cmp(const unsigned char *pat, const unsigned char *data, size_t len)
282 {
283 	size_t size;
284 	wchar_t *wdata, *wpat;
285 	unsigned int i;
286 
287 	if (iflag) {
288 		if ((size = mbstowcs(NULL, (const char *)data, 0)) ==
289 		    ((size_t) - 1))
290 			return (-1);
291 
292 		wdata = grep_malloc(size * sizeof(wint_t));
293 
294 		if (mbstowcs(wdata, (const char *)data, size) ==
295 		    ((size_t) - 1))
296 			return (-1);
297 
298 		if ((size = mbstowcs(NULL, (const char *)pat, 0)) ==
299 		    ((size_t) - 1))
300 			return (-1);
301 
302 		wpat = grep_malloc(size * sizeof(wint_t));
303 
304 		if (mbstowcs(wpat, (const char *)pat, size) == ((size_t) - 1))
305 			return (-1);
306 		for (i = 0; i < len; i++) {
307 			if ((towlower(wpat[i]) == towlower(wdata[i])) ||
308 			    ((grepbehave != GREP_FIXED) && wpat[i] == L'.'))
309 				continue;
310 			free(wpat);
311 			free(wdata);
312 				return (i);
313 		}
314 	} else {
315 		for (i = 0; i < len; i++) {
316 			if ((pat[i] == data[i]) || ((grepbehave != GREP_FIXED) &&
317 			    pat[i] == '.'))
318 				continue;
319 			return (i);
320 		}
321 	}
322 	return (-1);
323 }
324 
325 static inline void
grep_revstr(unsigned char * str,int len)326 grep_revstr(unsigned char *str, int len)
327 {
328 	int i;
329 	char c;
330 
331 	for (i = 0; i < len / 2; i++) {
332 		c = str[i];
333 		str[i] = str[len - i - 1];
334 		str[len - i - 1] = c;
335 	}
336 }
337