xref: /dragonfly/usr.bin/xargs/xargs.c (revision 9348a738)
1 /*-
2  * Copyright (c) 1990, 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  * John B. Roll Jr.
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. 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  * $xMach: xargs.c,v 1.6 2002/02/23 05:27:47 tim Exp $
33  *
34  * @(#) Copyright (c) 1990, 1993 The Regents of the University of California.  All rights reserved.
35  * @(#)xargs.c	8.1 (Berkeley) 6/6/93
36  * $FreeBSD: src/usr.bin/xargs/xargs.c,v 1.9.2.6 2003/06/01 21:40:35 mux Exp $
37  */
38 
39 #include <sys/types.h>
40 #include <sys/wait.h>
41 
42 #include <err.h>
43 #include <errno.h>
44 #include <fcntl.h>
45 #ifndef BOOTSTRAPPING
46 #include <langinfo.h>
47 #endif
48 #include <locale.h>
49 #include <paths.h>
50 #include <regex.h>
51 #include <stdio.h>
52 #include <stdlib.h>
53 #include <string.h>
54 #include <unistd.h>
55 
56 #include "pathnames.h"
57 
58 static void	parse_input(int, char *[]);
59 static void	prerun(int, char *[]);
60 static int	prompt(void);
61 static void	run(char **);
62 static void	usage(void);
63 void		strnsubst(char **, const char *, const char *, size_t);
64 static void	waitchildren(const char *, int);
65 
66 static char echo[] = _PATH_ECHO;
67 static char **av, **bxp, **ep, **expx, **xp;
68 static char *argp, *bbp, *ebp, *inpline, *p, *replstr;
69 static const char *eofstr;
70 static int count, insingle, indouble, oflag, pflag, tflag, Rflag, rval, zflag;
71 static int cnt, Iflag, jfound, Lflag, wasquoted, xflag;
72 static int curprocs, maxprocs;
73 static volatile int childerr;
74 
75 extern char **environ;
76 
77 int
78 main(int argc, char *argv[])
79 {
80 	char *tmp;
81 	long arg_max;
82 	int ch, Jflag, nargs, nflag, nline;
83 	size_t linelen;
84 
85 	inpline = replstr = NULL;
86 	ep = environ;
87 	eofstr = "";
88 	Jflag = nflag = 0;
89 
90 	(void)setlocale(LC_MESSAGES, "");
91 
92 	/*
93 	 * POSIX.2 limits the exec line length to ARG_MAX - 2K.  Running that
94 	 * caused some E2BIG errors, so it was changed to ARG_MAX - 4K.  Given
95 	 * that the smallest argument is 2 bytes in length, this means that
96 	 * the number of arguments is limited to:
97 	 *
98 	 *	 (ARG_MAX - 4K - LENGTH(utility + arguments)) / 2.
99 	 *
100 	 * We arbitrarily limit the number of arguments to 5000.  This is
101 	 * allowed by POSIX.2 as long as the resulting minimum exec line is
102 	 * at least LINE_MAX.  Realloc'ing as necessary is possible, but
103 	 * probably not worthwhile.
104 	 */
105 	nargs = 5000;
106 	if ((arg_max = sysconf(_SC_ARG_MAX)) == -1)
107 		errx(1, "sysconf(_SC_ARG_MAX) failed");
108 	nline = arg_max - 4 * 1024;
109 	while (*ep != NULL) {
110 		/* 1 byte for each '\0' */
111 		nline -= strlen(*ep++) + 1 + sizeof(*ep);
112 	}
113 
114 	maxprocs = 1;
115 	while ((ch = getopt(argc, argv, "0E:I:J:L:n:oprP:R:s:tx")) != -1)
116 		switch(ch) {
117 		case 'E':
118 			eofstr = optarg;
119 			break;
120 		case 'I':
121 			Jflag = 0;
122 			Iflag = 1;
123 			Lflag = 1;
124 			replstr = optarg;
125 			break;
126 		case 'J':
127 			Iflag = 0;
128 			Jflag = 1;
129 			replstr = optarg;
130 			break;
131 		case 'L':
132 			Lflag = strtol(optarg, &tmp, 10);
133 			if (*tmp != 0 || *optarg == 0)
134 				errx(1, "illegal argument count");
135 			break;
136 		case 'n':
137 			nflag = 1;
138 			if ((nargs = atoi(optarg)) <= 0)
139 				errx(1, "illegal argument count");
140 			break;
141 		case 'o':
142 			oflag = 1;
143 			break;
144 		case 'p':
145 			pflag = 1;
146 			break;
147 		case 'P':
148 			if ((maxprocs = atoi(optarg)) <= 0)
149 				errx(1, "max.processes must be >0");
150 			break;
151 		case 'R':
152 			if ((Rflag = atoi(optarg)) <= 0)
153 				errx(1, "illegal number of replacements");
154 			break;
155 		case 'r':
156 			/* GNU compatibility */
157 			/* --no-run-if-empty (this is our default) */
158 			break;
159 		case 's':
160 			nline = atoi(optarg);
161 			break;
162 		case 't':
163 			tflag = 1;
164 			break;
165 		case 'x':
166 			xflag = 1;
167 			break;
168 		case '0':
169 			zflag = 1;
170 			break;
171 		case '?':
172 		default:
173 			usage();
174 	}
175 	argc -= optind;
176 	argv += optind;
177 
178 	if (!Iflag && Rflag)
179 		usage();
180 	if (Iflag && !Rflag)
181 		Rflag = 5;
182 	if (xflag && !nflag)
183 		usage();
184 	if (Iflag || Lflag)
185 		xflag = 1;
186 	if (replstr != NULL && *replstr == '\0')
187 		errx(1, "replstr may not be empty");
188 
189 	/*
190 	 * Allocate pointers for the utility name, the utility arguments,
191 	 * the maximum arguments to be read from stdin and the trailing
192 	 * NULL.
193 	 */
194 	linelen = 1 + argc + nargs + 1;
195 	if ((av = bxp = malloc(linelen * sizeof(char **))) == NULL)
196 		errx(1, "malloc failed");
197 
198 	/*
199 	 * Use the user's name for the utility as argv[0], just like the
200 	 * shell.  Echo is the default.  Set up pointers for the user's
201 	 * arguments.
202 	 */
203 	if (*argv == NULL)
204 		cnt = strlen(*bxp++ = echo);
205 	else {
206 		do {
207 			if (Jflag && strcmp(*argv, replstr) == 0) {
208 				char **avj;
209 				jfound = 1;
210 				argv++;
211 				for (avj = argv; *avj; avj++)
212 					cnt += strlen(*avj) + 1;
213 				break;
214 			}
215 			cnt += strlen(*bxp++ = *argv) + 1;
216 		} while (*++argv != NULL);
217 	}
218 
219 	/*
220 	 * Set up begin/end/traversing pointers into the array.  The -n
221 	 * count doesn't include the trailing NULL pointer, so the malloc
222 	 * added in an extra slot.
223 	 */
224 	expx = (xp = bxp) + nargs;
225 
226 	/*
227 	 * Allocate buffer space for the arguments read from stdin and the
228 	 * trailing NULL.  Buffer space is defined as the default or specified
229 	 * space, minus the length of the utility name and arguments.  Set up
230 	 * begin/end/traversing pointers into the array.  The -s count does
231 	 * include the trailing NULL, so the malloc didn't add in an extra
232 	 * slot.
233 	 */
234 	nline -= cnt;
235 	if (nline <= 0)
236 		errx(1, "insufficient space for command");
237 
238 	if ((bbp = malloc((size_t)(nline + 1))) == NULL)
239 		errx(1, "malloc failed");
240 	ebp = (argp = p = bbp) + nline - 1;
241 	for (;;)
242 		parse_input(argc, argv);
243 }
244 
245 static void
246 parse_input(int argc, char *argv[])
247 {
248 	int ch, foundeof;
249 	char **avj;
250 
251 	foundeof = 0;
252 
253 	switch(ch = getchar()) {
254 	case EOF:
255 		/* No arguments since last exec. */
256 		if (p == bbp) {
257 			waitchildren(*argv, 1);
258 			exit(rval);
259 		}
260 		goto arg1;
261 	case ' ':
262 	case '\t':
263 		/* Quotes escape tabs and spaces. */
264 		if (insingle || indouble || zflag)
265 			goto addch;
266 		goto arg2;
267 	case '\0':
268 		if (zflag)
269 			goto arg2;
270 		goto addch;
271 	case '\n':
272 		count++;
273 		if (zflag)
274 			goto addch;
275 
276 		/* Quotes do not escape newlines. */
277 arg1:		if (insingle || indouble)
278 			errx(1, "unterminated quote");
279 arg2:
280 		foundeof = *eofstr != '\0' &&
281 		    strcmp(argp, eofstr) == 0;
282 
283 		/* Do not make empty args unless they are quoted */
284 		if ((argp != p || wasquoted) && !foundeof) {
285 			*p++ = '\0';
286 			*xp++ = argp;
287 			if (Iflag) {
288 				size_t curlen;
289 
290 				if (inpline == NULL)
291 					curlen = 0;
292 				else {
293 					/*
294 					 * If this string is not zero
295 					 * length, append a space for
296 					 * separation before the next
297 					 * argument.
298 					 */
299 					if ((curlen = strlen(inpline)))
300 						strcat(inpline, " ");
301 				}
302 				curlen++;
303 				/*
304 				 * Allocate enough to hold what we will
305 				 * be holding in a second, and to append
306 				 * a space next time through, if we have
307 				 * to.
308 				 */
309 				inpline = realloc(inpline, curlen + 2 +
310 				    strlen(argp));
311 				if (inpline == NULL)
312 					errx(1, "realloc failed");
313 				if (curlen == 1)
314 					strcpy(inpline, argp);
315 				else
316 					strcat(inpline, argp);
317 			}
318 		}
319 
320 		/*
321 		 * If max'd out on args or buffer, or reached EOF,
322 		 * run the command.  If xflag and max'd out on buffer
323 		 * but not on args, object.  Having reached the limit
324 		 * of input lines, as specified by -L is the same as
325 		 * maxing out on arguments.
326 		 */
327 		if (xp == expx || p > ebp || ch == EOF ||
328 		    (Lflag <= count && xflag) || foundeof) {
329 			if (xflag && xp != expx && p > ebp)
330 				errx(1, "insufficient space for arguments");
331 			if (jfound) {
332 				for (avj = argv; *avj; avj++)
333 					*xp++ = *avj;
334 			}
335 			prerun(argc, av);
336 			if (ch == EOF || foundeof) {
337 				waitchildren(*argv, 1);
338 				exit(rval);
339 			}
340 			p = bbp;
341 			xp = bxp;
342 			count = 0;
343 		}
344 		argp = p;
345 		wasquoted = 0;
346 		break;
347 	case '\'':
348 		if (indouble || zflag)
349 			goto addch;
350 		insingle = !insingle;
351 		wasquoted = 1;
352 		break;
353 	case '"':
354 		if (insingle || zflag)
355 			goto addch;
356 		indouble = !indouble;
357 		wasquoted = 1;
358 		break;
359 	case '\\':
360 		if (zflag)
361 			goto addch;
362 		/* Backslash escapes anything, is escaped by quotes. */
363 		if (!insingle && !indouble && (ch = getchar()) == EOF)
364 			errx(1, "backslash at EOF");
365 		/* FALLTHROUGH */
366 	default:
367 addch:		if (p < ebp) {
368 			*p++ = ch;
369 			break;
370 		}
371 
372 		/* If only one argument, not enough buffer space. */
373 		if (bxp == xp)
374 			errx(1, "insufficient space for argument");
375 		/* Didn't hit argument limit, so if xflag object. */
376 		if (xflag)
377 			errx(1, "insufficient space for arguments");
378 
379 		if (jfound) {
380 			for (avj = argv; *avj; avj++)
381 				*xp++ = *avj;
382 		}
383 		prerun(argc, av);
384 		xp = bxp;
385 		cnt = ebp - argp;
386 		memcpy(bbp, argp, (size_t)cnt);
387 		p = (argp = bbp) + cnt;
388 		*p++ = ch;
389 		break;
390 	}
391 	return;
392 }
393 
394 /*
395  * Do things necessary before run()'ing, such as -I substitution,
396  * and then call run().
397  */
398 static void
399 prerun(int argc, char *argv[])
400 {
401 	char **tmp, **tmp2, **avj;
402 	int repls;
403 
404 	repls = Rflag;
405 
406 	if (argc == 0 || repls == 0) {
407 		*xp = NULL;
408 		run(argv);
409 		return;
410 	}
411 
412 	avj = argv;
413 
414 	/*
415 	 * Allocate memory to hold the argument list, and
416 	 * a NULL at the tail.
417 	 */
418 	tmp = malloc((argc + 1) * sizeof(char**));
419 	if (tmp == NULL)
420 		errx(1, "malloc failed");
421 	tmp2 = tmp;
422 
423 	/*
424 	 * Save the first argument and iterate over it, we
425 	 * cannot do strnsubst() to it.
426 	 */
427 	if ((*tmp++ = strdup(*avj++)) == NULL)
428 		errx(1, "strdup failed");
429 
430 	/*
431 	 * For each argument to utility, if we have not used up
432 	 * the number of replacements we are allowed to do, and
433 	 * if the argument contains at least one occurrence of
434 	 * replstr, call strnsubst(), else just save the string.
435 	 * Iterations over elements of avj and tmp are done
436 	 * where appropriate.
437 	 */
438 	while (--argc) {
439 		*tmp = *avj++;
440 		if (repls && strstr(*tmp, replstr) != NULL) {
441 			strnsubst(tmp++, replstr, inpline, (size_t)255);
442 			repls--;
443 		} else {
444 			if ((*tmp = strdup(*tmp)) == NULL)
445 				errx(1, "strdup failed");
446 			tmp++;
447 		}
448 	}
449 
450 	/*
451 	 * Run it.
452 	 */
453 	*tmp = NULL;
454 	run(tmp2);
455 
456 	/*
457 	 * Walk from the tail to the head, free along the way.
458 	 */
459 	for (; tmp2 != tmp; tmp--)
460 		free(*tmp);
461 	/*
462 	 * Now free the list itself.
463 	 */
464 	free(tmp2);
465 
466 	/*
467 	 * Free the input line buffer, if we have one.
468 	 */
469 	if (inpline != NULL) {
470 		free(inpline);
471 		inpline = NULL;
472 	}
473 }
474 
475 static void
476 run(char **argv)
477 {
478 	char **avec;
479 	pid_t pid;
480 
481 	/*
482 	 * If the user wants to be notified of each command before it is
483 	 * executed, notify them.  If they want the notification to be
484 	 * followed by a prompt, then prompt them.
485 	 */
486 	if (tflag || pflag) {
487 		(void)fprintf(stderr, "%s", *argv);
488 		for (avec = argv + 1; *avec != NULL; ++avec)
489 			(void)fprintf(stderr, " %s", *avec);
490 		/*
491 		 * If the user has asked to be prompted, do so.
492 		 */
493 		if (pflag)
494 			/*
495 			 * If they asked not to exec, return without execution
496 			 * but if they asked to, go to the execution.  If we
497 			 * could not open their tty, break the switch and drop
498 			 * back to -t behaviour.
499 			 */
500 			switch (prompt()) {
501 			case 0:
502 				return;
503 			case 1:
504 				goto exec;
505 			case 2:
506 				break;
507 			}
508 		(void)fprintf(stderr, "\n");
509 		(void)fflush(stderr);
510 	}
511 exec:
512 	childerr = 0;
513 	switch(pid = vfork()) {
514 	case -1:
515 		err(1, "vfork");
516 	case 0:
517 		close(0);
518 		if (oflag) {
519 			if (open("/dev/tty", O_RDONLY) == -1)
520 				err(1, "open /dev/tty");
521 		} else {
522 			if (open("/dev/null", O_RDONLY) == -1)
523 				err(1, "open /dev/null");
524 		}
525 		execvp(argv[0], argv);
526 		childerr = errno;
527 		_exit(1);
528 	}
529 	curprocs++;
530 	waitchildren(*argv, 0);
531 }
532 
533 /*
534  * Handle child processes.
535  */
536 static void
537 waitchildren(const char *name, int waitall)
538 {
539 	pid_t pid;
540 	int status;
541 
542 	while ((pid = wait3(&status, !waitall && curprocs < maxprocs ?
543 		WNOHANG : 0, NULL)) > 0) {
544 		curprocs--;
545 
546 		/* If we couldn't invoke the utility, exit. */
547 		if (childerr != 0) {
548 			errno = childerr;
549 			err(errno == ENOENT ? 127 : 126, "%s", name);
550 		}
551 
552 		/*
553 		 * If utility signaled or exited with a value of 255,
554 		 * exit 1-125.
555 		 */
556 		if (WIFSIGNALED(status) || WEXITSTATUS(status) == 255)
557 			exit(1);
558 		if (WEXITSTATUS(status))
559 			rval = 1;
560         }
561 	if (pid == -1 && errno != ECHILD)
562 		err(1, "wait3");
563 }
564 
565 /*
566  * Prompt the user about running a command.
567  */
568 static int
569 prompt(void)
570 {
571 	regex_t cre;
572 	size_t rsize;
573 	int match;
574 	char *response;
575 	FILE *ttyfp;
576 
577 	if ((ttyfp = fopen(_PATH_TTY, "r")) == NULL)
578 		return (2);	/* Indicate that the TTY failed to open. */
579 	(void)fprintf(stderr, "?...");
580 	(void)fflush(stderr);
581 	if ((response = fgetln(ttyfp, &rsize)) == NULL ||
582 	    regcomp(&cre,
583 #ifdef BOOTSTRAPPING
584 		"^[yY]",
585 #else
586 		nl_langinfo(YESEXPR),
587 #endif
588 		REG_BASIC) != 0) {
589 		(void)fclose(ttyfp);
590 		return (0);
591 	}
592 	match = regexec(&cre, response, 0, NULL, 0);
593 	(void)fclose(ttyfp);
594 	regfree(&cre);
595 	return (match == 0);
596 }
597 
598 static void
599 usage(void)
600 {
601 	fprintf(stderr,
602 "usage: xargs [-0opt] [-E eofstr] [-I replstr [-R replacements]] [-J replstr]\n"
603 "             [-L number] [-n number [-x] [-P maxprocs] [-s size]\n"
604 "             [utility [argument ...]]\n");
605 	exit(1);
606 }
607