1 /* $NetBSD: look.c,v 1.13 2009/04/12 14:01:20 lukem Exp $ */ 2 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 #include <sys/cdefs.h> 36 #ifndef lint 37 __COPYRIGHT("@(#) Copyright (c) 1991, 1993\ 38 The Regents of the University of California. All rights reserved."); 39 #endif /* not lint */ 40 41 #ifndef lint 42 #if 0 43 static char sccsid[] = "@(#)look.c 8.2 (Berkeley) 5/4/95"; 44 #endif 45 __RCSID("$NetBSD: look.c,v 1.13 2009/04/12 14:01:20 lukem Exp $"); 46 #endif /* not lint */ 47 48 /* 49 * look -- find lines in a sorted list. 50 * 51 * The man page said that TABs and SPACEs participate in -d comparisons. 52 * In fact, they were ignored. This implements historic practice, not 53 * the manual page. 54 */ 55 56 #include <sys/types.h> 57 #include <sys/mman.h> 58 #include <sys/stat.h> 59 60 #include <ctype.h> 61 #include <errno.h> 62 #include <fcntl.h> 63 #include <limits.h> 64 #include <stdio.h> 65 #include <stdlib.h> 66 #include <string.h> 67 #include <unistd.h> 68 #include <err.h> 69 70 #include "pathnames.h" 71 72 /* 73 * FOLD and DICT convert characters to a normal form for comparison, 74 * according to the user specified flags. 75 * 76 * DICT expects integers because it uses a non-character value to 77 * indicate a character which should not participate in comparisons. 78 */ 79 #define EQUAL 0 80 #define GREATER 1 81 #define LESS (-1) 82 #define NO_COMPARE (-2) 83 84 #define FOLD(c) (isascii(c) && isupper(c) ? tolower(c) : (c)) 85 #define DICT(c) (isascii(c) && isalnum(c) ? (c) : NO_COMPARE) 86 87 int dflag, fflag; 88 89 char *binary_search __P((char *, char *, char *)); 90 int compare __P((char *, char *, char *)); 91 char *linear_search __P((char *, char *, char *)); 92 int look __P((char *, char *, char *)); 93 int main __P((int, char **)); 94 void print_from __P((char *, char *, char *)); 95 void usage __P((void)); 96 97 int 98 main(argc, argv) 99 int argc; 100 char *argv[]; 101 { 102 struct stat sb; 103 int ch, fd, termchar; 104 char *back, *front, *string, *p; 105 const char *file; 106 107 string = NULL; 108 file = _PATH_WORDS; 109 termchar = '\0'; 110 while ((ch = getopt(argc, argv, "dft:")) != -1) 111 switch(ch) { 112 case 'd': 113 dflag = 1; 114 break; 115 case 'f': 116 fflag = 1; 117 break; 118 case 't': 119 termchar = *optarg; 120 break; 121 case '?': 122 default: 123 usage(); 124 } 125 argc -= optind; 126 argv += optind; 127 128 switch (argc) { 129 case 2: /* Don't set -df for user. */ 130 string = *argv++; 131 file = *argv; 132 break; 133 case 1: /* But set -df by default. */ 134 dflag = fflag = 1; 135 string = *argv; 136 break; 137 default: 138 usage(); 139 } 140 141 if (termchar != '\0' && (p = strchr(string, termchar)) != NULL) 142 *++p = '\0'; 143 144 if ((fd = open(file, O_RDONLY, 0)) < 0 || fstat(fd, &sb)) 145 err(2, "%s", file); 146 if (sb.st_size > (off_t)SIZE_T_MAX) 147 err(2, "%s: %s", file, strerror(EFBIG)); 148 if ((front = mmap(NULL, (size_t)sb.st_size, 149 PROT_READ, MAP_FILE|MAP_SHARED, fd, (off_t)0)) == NULL) 150 err(2, "%s", file); 151 back = front + sb.st_size; 152 exit(look(string, front, back)); 153 } 154 155 int 156 look(string, front, back) 157 char *string, *front, *back; 158 { 159 int ch; 160 char *readp, *writep; 161 162 /* Reformat string string to avoid doing it multiple times later. */ 163 for (readp = writep = string; (ch = *readp++) != 0; ) { 164 if (fflag) 165 ch = FOLD(ch); 166 if (dflag) 167 ch = DICT(ch); 168 if (ch != NO_COMPARE) 169 *(writep++) = ch; 170 } 171 *writep = '\0'; 172 173 front = binary_search(string, front, back); 174 front = linear_search(string, front, back); 175 176 if (front) 177 print_from(string, front, back); 178 return (front ? 0 : 1); 179 } 180 181 182 /* 183 * Binary search for "string" in memory between "front" and "back". 184 * 185 * This routine is expected to return a pointer to the start of a line at 186 * *or before* the first word matching "string". Relaxing the constraint 187 * this way simplifies the algorithm. 188 * 189 * Invariants: 190 * front points to the beginning of a line at or before the first 191 * matching string. 192 * 193 * back points to the beginning of a line at or after the first 194 * matching line. 195 * 196 * Base of the Invariants. 197 * front = NULL; 198 * back = EOF; 199 * 200 * Advancing the Invariants: 201 * 202 * p = first newline after halfway point from front to back. 203 * 204 * If the string at "p" is not greater than the string to match, 205 * p is the new front. Otherwise it is the new back. 206 * 207 * Termination: 208 * 209 * The definition of the routine allows it return at any point, 210 * since front is always at or before the line to print. 211 * 212 * In fact, it returns when the chosen "p" equals "back". This 213 * implies that there exists a string is least half as long as 214 * (back - front), which in turn implies that a linear search will 215 * be no more expensive than the cost of simply printing a string or two. 216 * 217 * Trying to continue with binary search at this point would be 218 * more trouble than it's worth. 219 */ 220 #define SKIP_PAST_NEWLINE(p, back) \ 221 while (p < back && *p++ != '\n'); 222 223 char * 224 binary_search(string, front, back) 225 char *string, *front, *back; 226 { 227 char *p; 228 229 p = front + (back - front) / 2; 230 SKIP_PAST_NEWLINE(p, back); 231 232 /* 233 * If the file changes underneath us, make sure we don't 234 * infinitely loop. 235 */ 236 while (p < back && back > front) { 237 if (compare(string, p, back) == GREATER) 238 front = p; 239 else 240 back = p; 241 p = front + (back - front) / 2; 242 SKIP_PAST_NEWLINE(p, back); 243 } 244 return (front); 245 } 246 247 /* 248 * Find the first line that starts with string, linearly searching from front 249 * to back. 250 * 251 * Return NULL for no such line. 252 * 253 * This routine assumes: 254 * 255 * o front points at the first character in a line. 256 * o front is before or at the first line to be printed. 257 */ 258 char * 259 linear_search(string, front, back) 260 char *string, *front, *back; 261 { 262 while (front < back) { 263 switch (compare(string, front, back)) { 264 case EQUAL: /* Found it. */ 265 return (front); 266 break; 267 case LESS: /* No such string. */ 268 return (NULL); 269 break; 270 case GREATER: /* Keep going. */ 271 break; 272 } 273 SKIP_PAST_NEWLINE(front, back); 274 } 275 return (NULL); 276 } 277 278 /* 279 * Print as many lines as match string, starting at front. 280 */ 281 void 282 print_from(string, front, back) 283 char *string, *front, *back; 284 { 285 for (; front < back && compare(string, front, back) == EQUAL; ++front) { 286 for (; front < back && *front != '\n'; ++front) 287 if (putchar(*front) == EOF) 288 err(2, "stdout"); 289 if (putchar('\n') == EOF) 290 err(2, "stdout"); 291 } 292 } 293 294 /* 295 * Return LESS, GREATER, or EQUAL depending on how the string1 compares with 296 * string2 (s1 ??? s2). 297 * 298 * o Matches up to len(s1) are EQUAL. 299 * o Matches up to len(s2) are GREATER. 300 * 301 * Compare understands about the -f and -d flags, and treats comparisons 302 * appropriately. 303 * 304 * The string "s1" is null terminated. The string s2 is '\n' terminated (or 305 * "back" terminated). 306 */ 307 int 308 compare(s1, s2, back) 309 char *s1, *s2, *back; 310 { 311 int ch; 312 313 for (; *s1 && s2 < back && *s2 != '\n'; ++s1, ++s2) { 314 ch = *s2; 315 if (fflag) 316 ch = FOLD(ch); 317 if (dflag) 318 ch = DICT(ch); 319 320 if (ch == NO_COMPARE) { 321 ++s2; /* Ignore character in comparison. */ 322 continue; 323 } 324 if (*s1 != ch) 325 return (*s1 < ch ? LESS : GREATER); 326 } 327 return (*s1 ? GREATER : EQUAL); 328 } 329 330 void 331 usage() 332 { 333 (void)fprintf(stderr, "usage: look [-df] [-t char] string [file]\n"); 334 exit(2); 335 } 336