wordexp.c revision 198406
1/*-
2 * Copyright (c) 2002 Tim J. Robbins.
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 *
14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24 * SUCH DAMAGE.
25 */
26
27#include "namespace.h"
28#include <sys/cdefs.h>
29#include <sys/types.h>
30#include <sys/wait.h>
31#include <errno.h>
32#include <fcntl.h>
33#include <paths.h>
34#include <signal.h>
35#include <stdio.h>
36#include <stdlib.h>
37#include <string.h>
38#include <unistd.h>
39#include <wordexp.h>
40#include "un-namespace.h"
41
42__FBSDID("$FreeBSD: head/lib/libc/gen/wordexp.c 198406 2009-10-23 14:50:11Z jilles $");
43
44static int	we_askshell(const char *, wordexp_t *, int);
45static int	we_check(const char *, int);
46
47/*
48 * wordexp --
49 *	Perform shell word expansion on `words' and place the resulting list
50 *	of words in `we'. See wordexp(3).
51 *
52 *	Specified by IEEE Std. 1003.1-2001.
53 */
54int
55wordexp(const char * __restrict words, wordexp_t * __restrict we, int flags)
56{
57	int error;
58
59	if (flags & WRDE_REUSE)
60		wordfree(we);
61	if ((flags & WRDE_APPEND) == 0) {
62		we->we_wordc = 0;
63		we->we_wordv = NULL;
64		we->we_strings = NULL;
65		we->we_nbytes = 0;
66	}
67	if ((error = we_check(words, flags)) != 0) {
68		wordfree(we);
69		return (error);
70	}
71	if ((error = we_askshell(words, we, flags)) != 0) {
72		wordfree(we);
73		return (error);
74	}
75	return (0);
76}
77
78static size_t
79we_read_fully(int fd, char *buffer, size_t len)
80{
81	size_t done;
82	ssize_t nread;
83
84	done = 0;
85	do {
86		nread = _read(fd, buffer + done, len - done);
87		if (nread == -1 && errno == EINTR)
88			continue;
89		if (nread <= 0)
90			break;
91		done += nread;
92	} while (done != len);
93	return done;
94}
95
96/*
97 * we_askshell --
98 *	Use the `wordexp' /bin/sh builtin function to do most of the work
99 *	in expanding the word string. This function is complicated by
100 *	memory management.
101 */
102static int
103we_askshell(const char *words, wordexp_t *we, int flags)
104{
105	int pdes[2];			/* Pipe to child */
106	char bbuf[9];			/* Buffer for byte count */
107	char wbuf[9];			/* Buffer for word count */
108	long nwords, nbytes;		/* Number of words, bytes from child */
109	long i;				/* Handy integer */
110	size_t sofs;			/* Offset into we->we_strings */
111	size_t vofs;			/* Offset into we->we_wordv */
112	pid_t pid;			/* Process ID of child */
113	pid_t wpid;			/* waitpid return value */
114	int status;			/* Child exit status */
115	int error;			/* Our return value */
116	int serrno;			/* errno to return */
117	char *ifs;			/* IFS env. var. */
118	char *np, *p;			/* Handy pointers */
119	char *nstrings;			/* Temporary for realloc() */
120	char **nwv;			/* Temporary for realloc() */
121	sigset_t newsigblock, oldsigblock;
122
123	serrno = errno;
124	if ((ifs = getenv("IFS")) == NULL)
125		ifs = " \t\n";
126
127	if (pipe(pdes) < 0)
128		return (WRDE_NOSPACE);	/* XXX */
129	(void)sigemptyset(&newsigblock);
130	(void)sigaddset(&newsigblock, SIGCHLD);
131	(void)_sigprocmask(SIG_BLOCK, &newsigblock, &oldsigblock);
132	if ((pid = fork()) < 0) {
133		serrno = errno;
134		_close(pdes[0]);
135		_close(pdes[1]);
136		(void)_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
137		errno = serrno;
138		return (WRDE_NOSPACE);	/* XXX */
139	}
140	else if (pid == 0) {
141		/*
142		 * We are the child; just get /bin/sh to run the wordexp
143		 * builtin on `words'.
144		 */
145		int devnull;
146		char *cmd;
147
148		(void)_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
149		_close(pdes[0]);
150		if (_dup2(pdes[1], STDOUT_FILENO) < 0)
151			_exit(1);
152		_close(pdes[1]);
153		if (asprintf(&cmd, "wordexp%c%s\n", *ifs, words) < 0)
154			_exit(1);
155		if ((flags & WRDE_SHOWERR) == 0) {
156			if ((devnull = _open(_PATH_DEVNULL, O_RDWR, 0666)) < 0)
157				_exit(1);
158			if (_dup2(devnull, STDERR_FILENO) < 0)
159				_exit(1);
160			_close(devnull);
161		}
162		execl(_PATH_BSHELL, "sh", flags & WRDE_UNDEF ? "-u" : "+u",
163		    "-c", cmd, (char *)NULL);
164		_exit(1);
165	}
166
167	/*
168	 * We are the parent; read the output of the shell wordexp function,
169	 * which is a 32-bit hexadecimal word count, a 32-bit hexadecimal
170	 * byte count (not including terminating null bytes), followed by
171	 * the expanded words separated by nulls.
172	 */
173	_close(pdes[1]);
174	if (we_read_fully(pdes[0], wbuf, 8) != 8 ||
175			we_read_fully(pdes[0], bbuf, 8) != 8) {
176		error = flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX;
177		serrno = errno;
178		goto cleanup;
179	}
180	wbuf[8] = bbuf[8] = '\0';
181	nwords = strtol(wbuf, NULL, 16);
182	nbytes = strtol(bbuf, NULL, 16) + nwords;
183
184	/*
185	 * Allocate or reallocate (when flags & WRDE_APPEND) the word vector
186	 * and string storage buffers for the expanded words we're about to
187	 * read from the child.
188	 */
189	sofs = we->we_nbytes;
190	vofs = we->we_wordc;
191	if ((flags & (WRDE_DOOFFS|WRDE_APPEND)) == (WRDE_DOOFFS|WRDE_APPEND))
192		vofs += we->we_offs;
193	we->we_wordc += nwords;
194	we->we_nbytes += nbytes;
195	if ((nwv = realloc(we->we_wordv, (we->we_wordc + 1 +
196	    (flags & WRDE_DOOFFS ?  we->we_offs : 0)) *
197	    sizeof(char *))) == NULL) {
198		error = WRDE_NOSPACE;
199		goto cleanup;
200	}
201	we->we_wordv = nwv;
202	if ((nstrings = realloc(we->we_strings, we->we_nbytes)) == NULL) {
203		error = WRDE_NOSPACE;
204		goto cleanup;
205	}
206	for (i = 0; i < vofs; i++)
207		if (we->we_wordv[i] != NULL)
208			we->we_wordv[i] += nstrings - we->we_strings;
209	we->we_strings = nstrings;
210
211	if (we_read_fully(pdes[0], we->we_strings + sofs, nbytes) != nbytes) {
212		error = flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX;
213		serrno = errno;
214		goto cleanup;
215	}
216
217	error = 0;
218cleanup:
219	_close(pdes[0]);
220	do
221		wpid = _waitpid(pid, &status, 0);
222	while (wpid < 0 && errno == EINTR);
223	(void)_sigprocmask(SIG_SETMASK, &oldsigblock, NULL);
224	if (error != 0) {
225		errno = serrno;
226		return (error);
227	}
228	if (wpid < 0 || !WIFEXITED(status) || WEXITSTATUS(status) != 0)
229		return (flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX);
230
231	/*
232	 * Break the null-terminated expanded word strings out into
233	 * the vector.
234	 */
235	if (vofs == 0 && flags & WRDE_DOOFFS)
236		while (vofs < we->we_offs)
237			we->we_wordv[vofs++] = NULL;
238	p = we->we_strings + sofs;
239	while (nwords-- != 0) {
240		we->we_wordv[vofs++] = p;
241		if ((np = memchr(p, '\0', nbytes)) == NULL)
242			return (WRDE_NOSPACE);	/* XXX */
243		nbytes -= np - p + 1;
244		p = np + 1;
245	}
246	we->we_wordv[vofs] = NULL;
247
248	return (0);
249}
250
251/*
252 * we_check --
253 *	Check that the string contains none of the following unquoted
254 *	special characters: <newline> |&;<>(){}
255 *	or command substitutions when WRDE_NOCMD is set in flags.
256 */
257static int
258we_check(const char *words, int flags)
259{
260	char c;
261	int dquote, level, quote, squote;
262
263	quote = squote = dquote = 0;
264	while ((c = *words++) != '\0') {
265		switch (c) {
266		case '\\':
267			quote ^= 1;
268			continue;
269		case '\'':
270			if (quote + dquote == 0)
271				squote ^= 1;
272			break;
273		case '"':
274			if (quote + squote == 0)
275				dquote ^= 1;
276			break;
277		case '`':
278			if (quote + squote == 0 && flags & WRDE_NOCMD)
279				return (WRDE_CMDSUB);
280			while ((c = *words++) != '\0' && c != '`')
281				if (c == '\\' && (c = *words++) == '\0')
282					break;
283			if (c == '\0')
284				return (WRDE_SYNTAX);
285			break;
286		case '|': case '&': case ';': case '<': case '>':
287		case '{': case '}': case '(': case ')': case '\n':
288			if (quote + squote + dquote == 0)
289				return (WRDE_BADCHAR);
290			break;
291		case '$':
292			if ((c = *words++) == '\0')
293				break;
294			else if (quote + squote == 0 && c == '(') {
295				if (flags & WRDE_NOCMD && *words != '(')
296					return (WRDE_CMDSUB);
297				level = 1;
298				while ((c = *words++) != '\0') {
299					if (c == '\\') {
300						if ((c = *words++) == '\0')
301							break;
302					} else if (c == '(')
303						level++;
304					else if (c == ')' && --level == 0)
305						break;
306				}
307				if (c == '\0' || level != 0)
308					return (WRDE_SYNTAX);
309			} else if (quote + squote == 0 && c == '{') {
310				level = 1;
311				while ((c = *words++) != '\0') {
312					if (c == '\\') {
313						if ((c = *words++) == '\0')
314							break;
315					} else if (c == '{')
316						level++;
317					else if (c == '}' && --level == 0)
318						break;
319				}
320				if (c == '\0' || level != 0)
321					return (WRDE_SYNTAX);
322			} else
323				c = *--words;
324			break;
325		default:
326			break;
327		}
328		quote = 0;
329	}
330	if (quote + squote + dquote != 0)
331		return (WRDE_SYNTAX);
332
333	return (0);
334}
335
336/*
337 * wordfree --
338 *	Free the result of wordexp(). See wordexp(3).
339 *
340 *	Specified by IEEE Std. 1003.1-2001.
341 */
342void
343wordfree(wordexp_t *we)
344{
345
346	if (we == NULL)
347		return;
348	free(we->we_wordv);
349	free(we->we_strings);
350	we->we_wordv = NULL;
351	we->we_strings = NULL;
352	we->we_nbytes = 0;
353	we->we_wordc = 0;
354}
355