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