xref: /netbsd-src/lib/libc/gen/wordexp.c (revision 23c8222edbfb0f0932d88a8351d3a0cf817dfb9e)
1 /*	$NetBSD: wordexp.c,v 1.1 2004/07/13 15:42:03 seb Exp $	*/
2 
3 /*-
4  * Copyright (c) 2002 Tim J. Robbins.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26  * SUCH DAMAGE.
27  */
28 
29 #include "namespace.h"
30 #include <sys/cdefs.h>
31 #include <sys/types.h>
32 #include <assert.h>
33 #include <sys/wait.h>
34 #include <fcntl.h>
35 #include <paths.h>
36 #include <stdio.h>
37 #include <stdlib.h>
38 #include <string.h>
39 #include <unistd.h>
40 #include <wordexp.h>
41 
42 #if defined(LIBC_SCCS) && !defined(lint)
43 #if 0
44 __FBSDID("$FreeBSD: /repoman/r/ncvs/src/lib/libc/gen/wordexp.c,v 1.5 2004/04/09 11:32:32 tjr Exp $");
45 #else
46 __RCSID("$NetBSD: wordexp.c,v 1.1 2004/07/13 15:42:03 seb Exp $");
47 #endif
48 #endif /* LIBC_SCCS and not lint */
49 
50 static int	we_askshell(const char *, wordexp_t *, int);
51 static int	we_check(const char *, int);
52 
53 /*
54  * wordexp --
55  *	Perform shell word expansion on `words' and place the resulting list
56  *	of words in `we'. See wordexp(3).
57  *
58  */
59 int
60 wordexp(const char * __restrict words, wordexp_t * __restrict we, int flags)
61 {
62 	int error;
63 
64 	_DIAGASSERT(we != NULL);
65 	_DIAGASSERT(words != NULL);
66 	if (flags & WRDE_REUSE)
67 		wordfree(we);
68 	if ((flags & WRDE_APPEND) == 0) {
69 		we->we_wordc = 0;
70 		we->we_wordv = NULL;
71 		we->we_strings = NULL;
72 		we->we_nbytes = 0;
73 	}
74 	if ((error = we_check(words, flags)) != 0) {
75 		wordfree(we);
76 		return (error);
77 	}
78 	if ((error = we_askshell(words, we, flags)) != 0) {
79 		wordfree(we);
80 		return (error);
81 	}
82 	return (0);
83 }
84 
85 /*
86  * we_askshell --
87  *	Use the `wordexp' /bin/sh builtin function to do most of the work
88  *	in expanding the word string. This function is complicated by
89  *	memory management.
90  */
91 static int
92 we_askshell(const char *words, wordexp_t *we, int flags)
93 {
94 	int pdes[2];			/* Pipe to child */
95 	size_t nwords, nbytes;		/* Number of words, bytes from child */
96 	int i;				/* Handy integer */
97 	size_t sofs;			/* Offset into we->we_strings */
98 	size_t vofs;			/* Offset into we->we_wordv */
99 	pid_t pid;			/* Process ID of child */
100 	int status;			/* Child exit status */
101 	char *ifs;			/* IFS env. var. */
102 	char *np, *p;			/* Handy pointers */
103 	char *nstrings;			/* Temporary for realloc() */
104 	char **nwv;			/* Temporary for realloc() */
105 	FILE *fp;			/* Stream to read pipe */
106 	extern char **environ;
107 	char *cmd;
108 
109 	if ((ifs = getenv("IFS")) == NULL)
110 		ifs = " \t\n";
111 	if (asprintf(&cmd, "wordexp%c%s\n", *ifs, words) < 0)
112 		return (WRDE_NOSPACE);
113 	if (pipe(pdes) < 0) {
114 		free(cmd);
115 		return (WRDE_ERRNO);
116 	}
117 	if ((fp = fdopen(pdes[0], "r")) == NULL) {
118 		free(cmd);
119 		return (WRDE_ERRNO);
120 	}
121 	if ((pid = fork()) < 0) {
122 		free(cmd);
123 		fclose(fp);
124 		close(pdes[1]);
125 		return (WRDE_ERRNO);
126 	}
127 	else if (pid == 0) {
128 		/*
129 		 * We are the child; just get /bin/sh to run the wordexp
130 		 * builtin on `words'.
131 		 */
132 		int devnull;
133 
134 		close(pdes[0]);
135 		if (pdes[1] != STDOUT_FILENO) {
136 			if (dup2(pdes[1], STDOUT_FILENO) < 0)
137 				_exit(1);
138 			close(pdes[1]);
139 		}
140 		if ((flags & WRDE_SHOWERR) == 0) {
141 			if ((devnull = open(_PATH_DEVNULL, O_RDWR, 0666)) < 0)
142 				_exit(1);
143 			if (dup2(devnull, STDERR_FILENO) < 0)
144 				_exit(1);
145 			close(devnull);
146 		}
147 		execle(_PATH_BSHELL, "sh", flags & WRDE_UNDEF ? "-u" : "+u",
148 		    "-c", cmd, (char *)NULL, environ);
149 		_exit(1);
150 	}
151 
152 	/*
153 	 * We are the parent; read the output of the shell wordexp function,
154 	 * which is a decimal word count, an null, a decimal byte count,
155 	 * (not including terminating null bytes), a null and then followed
156 	 * by the expanded words separated by nulls.
157 	 */
158 	free(cmd);
159 	close(pdes[1]);
160 	/* read the word count */
161 	nwords = 0;
162 	while ((i = getc(fp)) != EOF) {
163 		if (i == '\0')
164 			break;
165 		nwords *= 10;
166 		nwords += (i - '0');
167 	}
168 	/* read the byte count */
169 	nbytes = 0;
170 	while ((i = getc(fp)) != EOF) {
171 		if (i == '\0')
172 			break;
173 		nbytes *= 10;
174 		nbytes += (i - '0');
175 	}
176 	if (i == EOF) {
177 		fclose(fp);
178 		waitpid(pid, &status, 0);
179 		return (flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX);
180 	}
181 	nbytes += nwords;
182 
183 	/*
184 	 * Allocate or reallocate (when flags & WRDE_APPEND) the word vector
185 	 * and string storage buffers for the expanded words we're about to
186 	 * read from the child.
187 	 */
188 	sofs = we->we_nbytes;
189 	vofs = we->we_wordc;
190 	if ((flags & (WRDE_DOOFFS|WRDE_APPEND)) == (WRDE_DOOFFS|WRDE_APPEND))
191 		vofs += we->we_offs;
192 	we->we_wordc += nwords;
193 	we->we_nbytes += nbytes;
194 	if ((nwv = realloc(we->we_wordv, (we->we_wordc + 1 +
195 	    (flags & WRDE_DOOFFS ?  we->we_offs : 0)) *
196 	    sizeof(char *))) == NULL) {
197 		fclose(fp);
198 		waitpid(pid, &status, 0);
199 		return (WRDE_NOSPACE);
200 	}
201 	we->we_wordv = nwv;
202 	if ((nstrings = realloc(we->we_strings, we->we_nbytes)) == NULL) {
203 		fclose(fp);
204 		waitpid(pid, &status, 0);
205 		return (WRDE_NOSPACE);
206 	}
207 	for (i = 0; i < vofs; i++)
208 		if (we->we_wordv[i] != NULL)
209 			we->we_wordv[i] += nstrings - we->we_strings;
210 	we->we_strings = nstrings;
211 
212 	if (fread(we->we_strings + sofs, sizeof(char), nbytes, fp) != nbytes) {
213 		fclose(fp);
214 		waitpid(pid, &status, 0);
215 		return (flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX);
216 	}
217 
218 	if (waitpid(pid, &status, 0) < 0 || !WIFEXITED(status) ||
219 	    WEXITSTATUS(status) != 0) {
220 		fclose(fp);
221 		return (flags & WRDE_UNDEF ? WRDE_BADVAL : WRDE_SYNTAX);
222 	}
223 	fclose(fp);
224 
225 	/*
226 	 * Break the null-terminated expanded word strings out into
227 	 * the vector.
228 	 */
229 	if (vofs == 0 && flags & WRDE_DOOFFS)
230 		while (vofs < we->we_offs)
231 			we->we_wordv[vofs++] = NULL;
232 	p = we->we_strings + sofs;
233 	while (nwords-- != 0) {
234 		we->we_wordv[vofs++] = p;
235 		if ((np = memchr(p, '\0', nbytes)) == NULL)
236 			return (WRDE_NOSPACE);	/* XXX */
237 		nbytes -= np - p + 1;
238 		p = np + 1;
239 	}
240 	we->we_wordv[vofs] = NULL;
241 
242 	return (0);
243 }
244 
245 /*
246  * we_check --
247  *	Check that the string contains none of the following unquoted
248  *	special characters: <newline> |&;<>(){}
249  *	or command substitutions when WRDE_NOCMD is set in flags.
250  */
251 static int
252 we_check(const char *words, int flags)
253 {
254 	char c;
255 	int dquote, level, quote, squote;
256 
257 	quote = squote = dquote = 0;
258 	while ((c = *words++) != '\0') {
259 		switch (c) {
260 		case '\\':
261 			quote ^= 1;
262 			continue;
263 		case '\'':
264 			if (quote + dquote == 0)
265 				squote ^= 1;
266 			break;
267 		case '"':
268 			if (quote + squote == 0)
269 				dquote ^= 1;
270 			break;
271 		case '`':
272 			if (quote + squote == 0 && flags & WRDE_NOCMD)
273 				return (WRDE_CMDSUB);
274 			while ((c = *words++) != '\0' && c != '`')
275 				if (c == '\\' && (c = *words++) == '\0')
276 					break;
277 			if (c == '\0')
278 				return (WRDE_SYNTAX);
279 			break;
280 		case '|': case '&': case ';': case '<': case '>':
281 		case '{': case '}': case '(': case ')': case '\n':
282 			if (quote + squote + dquote == 0)
283 				return (WRDE_BADCHAR);
284 			break;
285 		case '$':
286 			if ((c = *words++) == '\0')
287 				break;
288 			else if (quote + squote == 0 && c == '(') {
289 				if (flags & WRDE_NOCMD && *words != '(')
290 					return (WRDE_CMDSUB);
291 				level = 1;
292 				while ((c = *words++) != '\0') {
293 					if (c == '\\') {
294 						if ((c = *words++) == '\0')
295 							break;
296 					} else if (c == '(')
297 						level++;
298 					else if (c == ')' && --level == 0)
299 						break;
300 				}
301 				if (c == '\0' || level != 0)
302 					return (WRDE_SYNTAX);
303 			} else if (quote + squote == 0 && c == '{') {
304 				level = 1;
305 				while ((c = *words++) != '\0') {
306 					if (c == '\\') {
307 						if ((c = *words++) == '\0')
308 							break;
309 					} else if (c == '{')
310 						level++;
311 					else if (c == '}' && --level == 0)
312 						break;
313 				}
314 				if (c == '\0' || level != 0)
315 					return (WRDE_SYNTAX);
316 			} else
317 				c = *--words;
318 			break;
319 		default:
320 			break;
321 		}
322 		quote = 0;
323 	}
324 	if (quote + squote + dquote != 0)
325 		return (WRDE_SYNTAX);
326 
327 	return (0);
328 }
329 
330 /*
331  * wordfree --
332  *	Free the result of wordexp(). See wordexp(3).
333  *
334  */
335 void
336 wordfree(wordexp_t *we)
337 {
338 	_DIAGASSERT(we != NULL);
339 	free(we->we_wordv);
340 	free(we->we_strings);
341 	we->we_wordv = NULL;
342 	we->we_strings = NULL;
343 	we->we_nbytes = 0;
344 	we->we_wordc = 0;
345 }
346