1 /*
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1989, 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 * Adam S. Moskowitz of Menlo Consulting and Marciano Pitargue.
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 #ifndef lint
36 static const char copyright[] =
37 "@(#) Copyright (c) 1989, 1993\n\
38 The Regents of the University of California. All rights reserved.\n";
39 static const char sccsid[] = "@(#)cut.c 8.3 (Berkeley) 5/4/95";
40 #endif /* not lint */
41 #include <sys/cdefs.h>
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 static int wflag;
62
63 static size_t autostart, autostop, maxval;
64 static char * positions;
65
66 static int b_cut(FILE *, const char *);
67 static int b_n_cut(FILE *, const char *);
68 static int c_cut(FILE *, const char *);
69 static int f_cut(FILE *, const char *);
70 static void get_list(char *);
71 static int is_delim(wchar_t);
72 static void needpos(size_t);
73 static void usage(void);
74
75 int
main(int argc,char * argv[])76 main(int argc, char *argv[])
77 {
78 FILE *fp;
79 int (*fcn)(FILE *, const char *);
80 int ch, rval;
81 size_t n;
82
83 setlocale(LC_ALL, "");
84
85 fcn = NULL;
86 dchar = '\t'; /* default delimiter is \t */
87 strcpy(dcharmb, "\t");
88
89 while ((ch = getopt(argc, argv, "b:c:d:f:snw")) != -1)
90 switch(ch) {
91 case 'b':
92 get_list(optarg);
93 bflag = 1;
94 break;
95 case 'c':
96 get_list(optarg);
97 cflag = 1;
98 break;
99 case 'd':
100 n = mbrtowc(&dchar, optarg, MB_LEN_MAX, NULL);
101 if (dchar == '\0' || n != strlen(optarg))
102 errx(1, "bad delimiter");
103 strcpy(dcharmb, optarg);
104 dflag = 1;
105 break;
106 case 'f':
107 get_list(optarg);
108 fflag = 1;
109 break;
110 case 's':
111 sflag = 1;
112 break;
113 case 'n':
114 nflag = 1;
115 break;
116 case 'w':
117 wflag = 1;
118 break;
119 case '?':
120 default:
121 usage();
122 }
123 argc -= optind;
124 argv += optind;
125
126 if (fflag) {
127 if (bflag || cflag || nflag || (wflag && dflag))
128 usage();
129 } else if (!(bflag || cflag) || dflag || sflag || wflag)
130 usage();
131 else if (!bflag && nflag)
132 usage();
133
134 if (fflag)
135 fcn = f_cut;
136 else if (cflag)
137 fcn = MB_CUR_MAX > 1 ? c_cut : b_cut;
138 else if (bflag)
139 fcn = nflag && MB_CUR_MAX > 1 ? b_n_cut : b_cut;
140
141 rval = 0;
142 if (*argv)
143 for (; *argv; ++argv) {
144 if (strcmp(*argv, "-") == 0)
145 rval |= fcn(stdin, "stdin");
146 else {
147 if (!(fp = fopen(*argv, "r"))) {
148 warn("%s", *argv);
149 rval = 1;
150 continue;
151 }
152 fcn(fp, *argv);
153 (void)fclose(fp);
154 }
155 }
156 else
157 rval = fcn(stdin, "stdin");
158 exit(rval);
159 }
160
161 static void
get_list(char * list)162 get_list(char *list)
163 {
164 size_t setautostart, start, stop;
165 char *pos;
166 char *p;
167
168 /*
169 * set a byte in the positions array to indicate if a field or
170 * column is to be selected; use +1, it's 1-based, not 0-based.
171 * Numbers and number ranges may be overlapping, repeated, and in
172 * any order. We handle "-3-5" although there's no real reason to.
173 */
174 for (; (p = strsep(&list, ", \t")) != NULL;) {
175 setautostart = start = stop = 0;
176 if (*p == '-') {
177 ++p;
178 setautostart = 1;
179 }
180 if (isdigit((unsigned char)*p)) {
181 start = stop = strtol(p, &p, 10);
182 if (setautostart && start > autostart)
183 autostart = start;
184 }
185 if (*p == '-') {
186 if (isdigit((unsigned char)p[1]))
187 stop = strtol(p + 1, &p, 10);
188 if (*p == '-') {
189 ++p;
190 if (!autostop || autostop > stop)
191 autostop = stop;
192 }
193 }
194 if (*p)
195 errx(1, "[-bcf] list: illegal list value");
196 if (!stop || !start)
197 errx(1, "[-bcf] list: values may not include zero");
198 if (maxval < stop) {
199 maxval = stop;
200 needpos(maxval + 1);
201 }
202 for (pos = positions + start; start++ <= stop; *pos++ = 1);
203 }
204
205 /* overlapping ranges */
206 if (autostop && maxval > autostop) {
207 maxval = autostop;
208 needpos(maxval + 1);
209 }
210
211 /* reversed range with autostart */
212 if (maxval < autostart) {
213 maxval = autostart;
214 needpos(maxval + 1);
215 }
216
217 /* set autostart */
218 if (autostart)
219 memset(positions + 1, '1', autostart);
220 }
221
222 static void
needpos(size_t n)223 needpos(size_t n)
224 {
225 static size_t npos;
226 size_t oldnpos;
227
228 /* Grow the positions array to at least the specified size. */
229 if (n > npos) {
230 oldnpos = npos;
231 if (npos == 0)
232 npos = n;
233 while (n > npos)
234 npos *= 2;
235 if ((positions = realloc(positions, npos)) == NULL)
236 err(1, "realloc");
237 memset((char *)positions + oldnpos, 0, npos - oldnpos);
238 }
239 }
240
241 static int
b_cut(FILE * fp,const char * fname __unused)242 b_cut(FILE *fp, const char *fname __unused)
243 {
244 int ch, col;
245 char *pos;
246
247 ch = 0;
248 for (;;) {
249 pos = positions + 1;
250 for (col = maxval; col; --col) {
251 if ((ch = getc(fp)) == EOF)
252 return (0);
253 if (ch == '\n')
254 break;
255 if (*pos++)
256 (void)putchar(ch);
257 }
258 if (ch != '\n') {
259 if (autostop)
260 while ((ch = getc(fp)) != EOF && ch != '\n')
261 (void)putchar(ch);
262 else
263 while ((ch = getc(fp)) != EOF && ch != '\n');
264 }
265 (void)putchar('\n');
266 }
267 return (0);
268 }
269
270 /*
271 * Cut based on byte positions, taking care not to split multibyte characters.
272 * Although this function also handles the case where -n is not specified,
273 * b_cut() ought to be much faster.
274 */
275 static int
b_n_cut(FILE * fp,const char * fname)276 b_n_cut(FILE *fp, const char *fname)
277 {
278 size_t col, i, lbuflen;
279 char *lbuf;
280 int canwrite, clen, warned;
281 mbstate_t mbs;
282
283 memset(&mbs, 0, sizeof(mbs));
284 warned = 0;
285 while ((lbuf = fgetln(fp, &lbuflen)) != NULL) {
286 for (col = 0; lbuflen > 0; col += clen) {
287 if ((clen = mbrlen(lbuf, lbuflen, &mbs)) < 0) {
288 if (!warned) {
289 warn("%s", fname);
290 warned = 1;
291 }
292 memset(&mbs, 0, sizeof(mbs));
293 clen = 1;
294 }
295 if (clen == 0 || *lbuf == '\n')
296 break;
297 if (col < maxval && !positions[1 + col]) {
298 /*
299 * Print the character if (1) after an initial
300 * segment of un-selected bytes, the rest of
301 * it is selected, and (2) the last byte is
302 * selected.
303 */
304 i = col;
305 while (i < col + clen && i < maxval &&
306 !positions[1 + i])
307 i++;
308 canwrite = i < col + clen;
309 for (; i < col + clen && i < maxval; i++)
310 canwrite &= positions[1 + i];
311 if (canwrite)
312 fwrite(lbuf, 1, clen, stdout);
313 } else {
314 /*
315 * Print the character if all of it has
316 * been selected.
317 */
318 canwrite = 1;
319 for (i = col; i < col + clen; i++)
320 if ((i >= maxval && !autostop) ||
321 (i < maxval && !positions[1 + i])) {
322 canwrite = 0;
323 break;
324 }
325 if (canwrite)
326 fwrite(lbuf, 1, clen, stdout);
327 }
328 lbuf += clen;
329 lbuflen -= clen;
330 }
331 if (lbuflen > 0)
332 putchar('\n');
333 }
334 return (warned);
335 }
336
337 static int
c_cut(FILE * fp,const char * fname)338 c_cut(FILE *fp, const char *fname)
339 {
340 wint_t ch;
341 int col;
342 char *pos;
343
344 ch = 0;
345 for (;;) {
346 pos = positions + 1;
347 for (col = maxval; col; --col) {
348 if ((ch = getwc(fp)) == WEOF)
349 goto out;
350 if (ch == '\n')
351 break;
352 if (*pos++)
353 (void)putwchar(ch);
354 }
355 if (ch != '\n') {
356 if (autostop)
357 while ((ch = getwc(fp)) != WEOF && ch != '\n')
358 (void)putwchar(ch);
359 else
360 while ((ch = getwc(fp)) != WEOF && ch != '\n');
361 }
362 (void)putwchar('\n');
363 }
364 out:
365 if (ferror(fp)) {
366 warn("%s", fname);
367 return (1);
368 }
369 return (0);
370 }
371
372 static int
is_delim(wchar_t ch)373 is_delim(wchar_t ch)
374 {
375 if (wflag) {
376 if (ch == ' ' || ch == '\t')
377 return 1;
378 } else {
379 if (ch == dchar)
380 return 1;
381 }
382 return 0;
383 }
384
385 static int
f_cut(FILE * fp,const char * fname)386 f_cut(FILE *fp, const char *fname)
387 {
388 wchar_t ch;
389 int field, i, isdelim;
390 char *pos, *p;
391 int output;
392 char *lbuf, *mlbuf;
393 size_t clen, lbuflen, reallen;
394
395 mlbuf = NULL;
396 while ((lbuf = fgetln(fp, &lbuflen)) != NULL) {
397 reallen = lbuflen;
398 /* Assert EOL has a newline. */
399 if (*(lbuf + lbuflen - 1) != '\n') {
400 /* Can't have > 1 line with no trailing newline. */
401 mlbuf = malloc(lbuflen + 1);
402 if (mlbuf == NULL)
403 err(1, "malloc");
404 memcpy(mlbuf, lbuf, lbuflen);
405 *(mlbuf + lbuflen) = '\n';
406 lbuf = mlbuf;
407 reallen++;
408 }
409 output = 0;
410 for (isdelim = 0, p = lbuf;; p += clen) {
411 clen = mbrtowc(&ch, p, lbuf + reallen - p, NULL);
412 if (clen == (size_t)-1 || clen == (size_t)-2) {
413 warnc(EILSEQ, "%s", fname);
414 free(mlbuf);
415 return (1);
416 }
417 if (clen == 0)
418 clen = 1;
419 /* this should work if newline is delimiter */
420 if (is_delim(ch))
421 isdelim = 1;
422 if (ch == '\n') {
423 if (!isdelim && !sflag)
424 (void)fwrite(lbuf, lbuflen, 1, stdout);
425 break;
426 }
427 }
428 if (!isdelim)
429 continue;
430
431 pos = positions + 1;
432 for (field = maxval, p = lbuf; field; --field, ++pos) {
433 if (*pos && output++)
434 for (i = 0; dcharmb[i] != '\0'; i++)
435 putchar(dcharmb[i]);
436 for (;;) {
437 clen = mbrtowc(&ch, p, lbuf + reallen - p,
438 NULL);
439 if (clen == (size_t)-1 || clen == (size_t)-2) {
440 warnc(EILSEQ, "%s", fname);
441 free(mlbuf);
442 return (1);
443 }
444 if (clen == 0)
445 clen = 1;
446 p += clen;
447 if (ch == '\n' || is_delim(ch)) {
448 /* compress whitespace */
449 if (wflag && ch != '\n')
450 while (is_delim(*p))
451 p++;
452 break;
453 }
454 if (*pos)
455 for (i = 0; i < (int)clen; i++)
456 putchar(p[i - clen]);
457 }
458 if (ch == '\n')
459 break;
460 }
461 if (ch != '\n') {
462 if (autostop) {
463 if (output)
464 for (i = 0; dcharmb[i] != '\0'; i++)
465 putchar(dcharmb[i]);
466 for (; (ch = *p) != '\n'; ++p)
467 (void)putchar(ch);
468 } else
469 for (; (ch = *p) != '\n'; ++p);
470 }
471 (void)putchar('\n');
472 }
473 free(mlbuf);
474 return (0);
475 }
476
477 static void
usage(void)478 usage(void)
479 {
480 (void)fprintf(stderr, "%s\n%s\n%s\n",
481 "usage: cut -b list [-n] [file ...]",
482 " cut -c list [file ...]",
483 " cut -f list [-s] [-w | -d delim] [file ...]");
484 exit(1);
485 }
486