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