1 /* 2 * Copyright (c) 1989, 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 * Adam S. Moskowitz of Menlo Consulting and Marciano Pitargue. 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) 1989, 1993\n\ 36 The Regents of the University of California. All rights reserved.\n"; 37 static const char sccsid[] = "@(#)cut.c 8.3 (Berkeley) 5/4/95"; 38 #endif /* not lint */ 39 #include <sys/cdefs.h> 40 __FBSDID("$FreeBSD$"); 41 42 #include <ctype.h> 43 #include <err.h> 44 #include <errno.h> 45 #include <limits.h> 46 #include <locale.h> 47 #include <stdio.h> 48 #include <stdlib.h> 49 #include <string.h> 50 #include <unistd.h> 51 #include <wchar.h> 52 53 static int bflag; 54 static int cflag; 55 static wchar_t dchar; 56 static char dcharmb[MB_LEN_MAX + 1]; 57 static int dflag; 58 static int fflag; 59 static int nflag; 60 static int sflag; 61 62 static size_t autostart, autostop, maxval; 63 static char * positions; 64 65 static int b_cut(FILE *, const char *); 66 static int b_n_cut(FILE *, const char *); 67 static int c_cut(FILE *, const char *); 68 static int f_cut(FILE *, const char *); 69 static void get_list(char *); 70 static void needpos(size_t); 71 static void usage(void); 72 73 int 74 main(int argc, char *argv[]) 75 { 76 FILE *fp; 77 int (*fcn)(FILE *, const char *); 78 int ch, rval; 79 size_t n; 80 81 setlocale(LC_ALL, ""); 82 83 fcn = NULL; 84 dchar = '\t'; /* default delimiter is \t */ 85 strcpy(dcharmb, "\t"); 86 87 while ((ch = getopt(argc, argv, "b:c:d:f:sn")) != -1) 88 switch(ch) { 89 case 'b': 90 get_list(optarg); 91 bflag = 1; 92 break; 93 case 'c': 94 get_list(optarg); 95 cflag = 1; 96 break; 97 case 'd': 98 n = mbrtowc(&dchar, optarg, MB_LEN_MAX, NULL); 99 if (dchar == '\0' || n != strlen(optarg)) 100 errx(1, "bad delimiter"); 101 strcpy(dcharmb, optarg); 102 dflag = 1; 103 break; 104 case 'f': 105 get_list(optarg); 106 fflag = 1; 107 break; 108 case 's': 109 sflag = 1; 110 break; 111 case 'n': 112 nflag = 1; 113 break; 114 case '?': 115 default: 116 usage(); 117 } 118 argc -= optind; 119 argv += optind; 120 121 if (fflag) { 122 if (bflag || cflag || nflag) 123 usage(); 124 } else if (!(bflag || cflag) || dflag || sflag) 125 usage(); 126 else if (!bflag && nflag) 127 usage(); 128 129 if (fflag) 130 fcn = f_cut; 131 else if (cflag) 132 fcn = MB_CUR_MAX > 1 ? c_cut : b_cut; 133 else if (bflag) 134 fcn = nflag && MB_CUR_MAX > 1 ? b_n_cut : b_cut; 135 136 rval = 0; 137 if (*argv) 138 for (; *argv; ++argv) { 139 if (strcmp(*argv, "-") == 0) 140 rval |= fcn(stdin, "stdin"); 141 else { 142 if (!(fp = fopen(*argv, "r"))) { 143 warn("%s", *argv); 144 rval = 1; 145 continue; 146 } 147 fcn(fp, *argv); 148 (void)fclose(fp); 149 } 150 } 151 else 152 rval = fcn(stdin, "stdin"); 153 exit(rval); 154 } 155 156 static void 157 get_list(char *list) 158 { 159 size_t setautostart, start, stop; 160 char *pos; 161 char *p; 162 163 /* 164 * set a byte in the positions array to indicate if a field or 165 * column is to be selected; use +1, it's 1-based, not 0-based. 166 * Numbers and number ranges may be overlapping, repeated, and in 167 * any order. We handle "-3-5" although there's no real reason too. 168 */ 169 for (; (p = strsep(&list, ", \t")) != NULL;) { 170 setautostart = start = stop = 0; 171 if (*p == '-') { 172 ++p; 173 setautostart = 1; 174 } 175 if (isdigit((unsigned char)*p)) { 176 start = stop = strtol(p, &p, 10); 177 if (setautostart && start > autostart) 178 autostart = start; 179 } 180 if (*p == '-') { 181 if (isdigit((unsigned char)p[1])) 182 stop = strtol(p + 1, &p, 10); 183 if (*p == '-') { 184 ++p; 185 if (!autostop || autostop > stop) 186 autostop = stop; 187 } 188 } 189 if (*p) 190 errx(1, "[-cf] list: illegal list value"); 191 if (!stop || !start) 192 errx(1, "[-cf] list: values may not include zero"); 193 if (maxval < stop) { 194 maxval = stop; 195 needpos(maxval + 1); 196 } 197 for (pos = positions + start; start++ <= stop; *pos++ = 1); 198 } 199 200 /* overlapping ranges */ 201 if (autostop && maxval > autostop) { 202 maxval = autostop; 203 needpos(maxval + 1); 204 } 205 206 /* set autostart */ 207 if (autostart) 208 memset(positions + 1, '1', autostart); 209 } 210 211 static void 212 needpos(size_t n) 213 { 214 static size_t npos; 215 size_t oldnpos; 216 217 /* Grow the positions array to at least the specified size. */ 218 if (n > npos) { 219 oldnpos = npos; 220 if (npos == 0) 221 npos = n; 222 while (n > npos) 223 npos *= 2; 224 if ((positions = realloc(positions, npos)) == NULL) 225 err(1, "realloc"); 226 memset((char *)positions + oldnpos, 0, npos - oldnpos); 227 } 228 } 229 230 static int 231 b_cut(FILE *fp, const char *fname __unused) 232 { 233 int ch, col; 234 char *pos; 235 236 ch = 0; 237 for (;;) { 238 pos = positions + 1; 239 for (col = maxval; col; --col) { 240 if ((ch = getc(fp)) == EOF) 241 return (0); 242 if (ch == '\n') 243 break; 244 if (*pos++) 245 (void)putchar(ch); 246 } 247 if (ch != '\n') { 248 if (autostop) 249 while ((ch = getc(fp)) != EOF && ch != '\n') 250 (void)putchar(ch); 251 else 252 while ((ch = getc(fp)) != EOF && ch != '\n'); 253 } 254 (void)putchar('\n'); 255 } 256 return (0); 257 } 258 259 /* 260 * Cut based on byte positions, taking care not to split multibyte characters. 261 * Although this function also handles the case where -n is not specified, 262 * b_cut() ought to be much faster. 263 */ 264 static int 265 b_n_cut(FILE *fp, const char *fname) 266 { 267 size_t col, i, lbuflen; 268 char *lbuf; 269 int canwrite, clen, warned; 270 mbstate_t mbs; 271 272 memset(&mbs, 0, sizeof(mbs)); 273 warned = 0; 274 while ((lbuf = fgetln(fp, &lbuflen)) != NULL) { 275 for (col = 0; lbuflen > 0; col += clen) { 276 if ((clen = mbrlen(lbuf, lbuflen, &mbs)) < 0) { 277 if (!warned) { 278 warn("%s", fname); 279 warned = 1; 280 } 281 memset(&mbs, 0, sizeof(mbs)); 282 clen = 1; 283 } 284 if (clen == 0 || *lbuf == '\n') 285 break; 286 if (col < maxval && !positions[1 + col]) { 287 /* 288 * Print the character if (1) after an initial 289 * segment of un-selected bytes, the rest of 290 * it is selected, and (2) the last byte is 291 * selected. 292 */ 293 i = col; 294 while (i < col + clen && i < maxval && 295 !positions[1 + i]) 296 i++; 297 canwrite = i < col + clen; 298 for (; i < col + clen && i < maxval; i++) 299 canwrite &= positions[1 + i]; 300 if (canwrite) 301 fwrite(lbuf, 1, clen, stdout); 302 } else { 303 /* 304 * Print the character if all of it has 305 * been selected. 306 */ 307 canwrite = 1; 308 for (i = col; i < col + clen; i++) 309 if ((i >= maxval && !autostop) || 310 (i < maxval && !positions[1 + i])) { 311 canwrite = 0; 312 break; 313 } 314 if (canwrite) 315 fwrite(lbuf, 1, clen, stdout); 316 } 317 lbuf += clen; 318 lbuflen -= clen; 319 } 320 if (lbuflen > 0) 321 putchar('\n'); 322 } 323 return (warned); 324 } 325 326 static int 327 c_cut(FILE *fp, const char *fname) 328 { 329 wint_t ch; 330 int col; 331 char *pos; 332 333 ch = 0; 334 for (;;) { 335 pos = positions + 1; 336 for (col = maxval; col; --col) { 337 if ((ch = getwc(fp)) == WEOF) 338 goto out; 339 if (ch == '\n') 340 break; 341 if (*pos++) 342 (void)putwchar(ch); 343 } 344 if (ch != '\n') { 345 if (autostop) 346 while ((ch = getwc(fp)) != WEOF && ch != '\n') 347 (void)putwchar(ch); 348 else 349 while ((ch = getwc(fp)) != WEOF && ch != '\n'); 350 } 351 (void)putwchar('\n'); 352 } 353 out: 354 if (ferror(fp)) { 355 warn("%s", fname); 356 return (1); 357 } 358 return (0); 359 } 360 361 static int 362 f_cut(FILE *fp, const char *fname) 363 { 364 wchar_t ch; 365 int field, i, isdelim; 366 char *pos, *p; 367 wchar_t sep; 368 int output; 369 char *lbuf, *mlbuf; 370 size_t clen, lbuflen, reallen; 371 372 mlbuf = NULL; 373 for (sep = dchar; (lbuf = fgetln(fp, &lbuflen)) != NULL;) { 374 reallen = lbuflen; 375 /* Assert EOL has a newline. */ 376 if (*(lbuf + lbuflen - 1) != '\n') { 377 /* Can't have > 1 line with no trailing newline. */ 378 mlbuf = malloc(lbuflen + 1); 379 if (mlbuf == NULL) 380 err(1, "malloc"); 381 memcpy(mlbuf, lbuf, lbuflen); 382 *(mlbuf + lbuflen) = '\n'; 383 lbuf = mlbuf; 384 reallen++; 385 } 386 output = 0; 387 for (isdelim = 0, p = lbuf;; p += clen) { 388 clen = mbrtowc(&ch, p, lbuf + reallen - p, NULL); 389 if (clen == (size_t)-1 || clen == (size_t)-2) { 390 warnc(EILSEQ, "%s", fname); 391 free(mlbuf); 392 return (1); 393 } 394 if (clen == 0) 395 clen = 1; 396 /* this should work if newline is delimiter */ 397 if (ch == sep) 398 isdelim = 1; 399 if (ch == '\n') { 400 if (!isdelim && !sflag) 401 (void)fwrite(lbuf, lbuflen, 1, stdout); 402 break; 403 } 404 } 405 if (!isdelim) 406 continue; 407 408 pos = positions + 1; 409 for (field = maxval, p = lbuf; field; --field, ++pos) { 410 if (*pos && output++) 411 for (i = 0; dcharmb[i] != '\0'; i++) 412 putchar(dcharmb[i]); 413 for (;;) { 414 clen = mbrtowc(&ch, p, lbuf + reallen - p, 415 NULL); 416 if (clen == (size_t)-1 || clen == (size_t)-2) { 417 warnc(EILSEQ, "%s", fname); 418 free(mlbuf); 419 return (1); 420 } 421 if (clen == 0) 422 clen = 1; 423 p += clen; 424 if (ch == '\n' || ch == sep) 425 break; 426 if (*pos) 427 for (i = 0; i < (int)clen; i++) 428 putchar(p[i - clen]); 429 } 430 if (ch == '\n') 431 break; 432 } 433 if (ch != '\n') { 434 if (autostop) { 435 if (output) 436 for (i = 0; dcharmb[i] != '\0'; i++) 437 putchar(dcharmb[i]); 438 for (; (ch = *p) != '\n'; ++p) 439 (void)putchar(ch); 440 } else 441 for (; (ch = *p) != '\n'; ++p); 442 } 443 (void)putchar('\n'); 444 } 445 free(mlbuf); 446 return (0); 447 } 448 449 static void 450 usage(void) 451 { 452 (void)fprintf(stderr, "%s\n%s\n%s\n", 453 "usage: cut -b list [-n] [file ...]", 454 " cut -c list [file ...]", 455 " cut -f list [-s] [-d delim] [file ...]"); 456 exit(1); 457 } 458