glob.c revision d435b43c338a692531af0fe3df29cb968a863403 
1/*
2 * This file is part of ltrace.
3 * Copyright (C) 2007,2008,2012,2013 Petr Machata, Red Hat Inc.
4 *
5 * This program is free software; you can redistribute it and/or
6 * modify it under the terms of the GNU General Public License as
7 * published by the Free Software Foundation; either version 2 of the
8 * License, or (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful, but
11 * WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
13 * General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
18 * 02110-1301 USA
19 */
20
21#include <sys/types.h>
22#include <regex.h>
23#include <string.h>
24#include <stdlib.h>
25#include <assert.h>
26
27static ssize_t
28match_character_class(const char *glob, size_t length, size_t from)
29{
30	size_t i = 0;
31	if (length > 0)
32		for (i = from + 2; i < length - 1 && glob[++i] != ':'; )
33			;
34	if (i >= length || glob[++i] != ']')
35		return -1;
36	return i;
37}
38
39static ssize_t
40match_brack(const char *glob, size_t length, size_t from, int *exclmp)
41{
42	size_t i = from + 1;
43
44	if (i >= length)
45		return -1;
46
47	/* Complement operator.  */
48	*exclmp = 0;
49	if (glob[i] == '^' || glob[i] == '!') {
50		*exclmp = glob[i++] == '!';
51		if (i >= length)
52			return -1;
53	}
54
55	/* On first character, both [ and ] are legal.  But when [ is
56	 * followed with :, it's character class.  */
57	if (glob[i] == '[' && glob[i + 1] == ':') {
58		ssize_t j = match_character_class(glob, length, i);
59		if (j < 0)
60		fail:
61			return -1;
62		i = j;
63	}
64	++i; /* skip any character, including [ or ]  */
65
66	for (; i < length; ++i) {
67		char c = glob[i];
68		if (c == '[' && glob[i + 1] == ':') {
69			ssize_t j = match_character_class(glob, length, i);
70			if (j < 0)
71				goto fail;
72			i = j;
73
74		} else if (c == ']') {
75			return i;
76		}
77	}
78	return -1;
79}
80
81static int
82append(char **bufp, const char *str, size_t str_size,
83       size_t *sizep, size_t *allocp)
84{
85	if (str_size == 0)
86		str_size = strlen(str);
87	size_t nsize = *sizep + str_size;
88	if (nsize > *allocp) {
89		size_t nalloc = nsize * 2;
90		char *nbuf = realloc(*bufp, nalloc);
91		if (nbuf == NULL)
92			return -1;
93		*allocp = nalloc;
94		*bufp = nbuf;
95	}
96
97	memcpy(*bufp + *sizep, str, str_size);
98	*sizep = nsize;
99	return 0;
100}
101
102static int
103glob_to_regex(const char *glob, char **retp)
104{
105	size_t allocd = 0;
106	size_t size = 0;
107	char *buf = NULL;
108
109	size_t length = strlen(glob);
110	int escape = 0;
111	size_t i;
112	for(i = 0; i < length; ++i) {
113		char c = glob[i];
114		if (escape) {
115			if (c == '\\') {
116				if (append(&buf, "\\\\", 0,
117					   &size, &allocd) < 0) {
118				fail:
119					free(buf);
120					return REG_ESPACE;
121				}
122
123			} else if (c == '*') {
124				if (append(&buf, "\\*", 0, &size, &allocd) < 0)
125					goto fail;
126			} else if (c == '?') {
127				if (append(&buf, "?", 0, &size, &allocd) < 0)
128					goto fail;
129			} else if (append(&buf, (char[]){ '\\', c }, 2,
130					  &size, &allocd) < 0)
131				goto fail;
132			escape = 0;
133		} else {
134			if (c == '\\')
135				escape = 1;
136			else if (c == '[') {
137				int exclm;
138				ssize_t j = match_brack(glob, length, i, &exclm);
139				if (j < 0) {
140					free(buf);
141					return REG_EBRACK;
142				}
143				if (exclm
144				    && append(&buf, "[^", 2,
145					      &size, &allocd) < 0)
146					goto fail;
147				if (append(&buf, glob + i + 2*exclm,
148					   j - i + 1 - 2*exclm,
149					   &size, &allocd) < 0)
150					goto fail;
151				i = j;
152
153			} else if (c == '*') {
154				if (append(&buf, ".*", 0, &size, &allocd) < 0)
155					goto fail;
156			} else if (c == '?') {
157				if (append(&buf, ".", 0, &size, &allocd) < 0)
158					goto fail;
159			} else if (c == '.') {
160				if (append(&buf, "\\.", 0, &size, &allocd) < 0)
161					goto fail;
162			} else if (append(&buf, &c, 1, &size, &allocd) < 0)
163				goto fail;
164		}
165	}
166
167	if (escape) {
168		free(buf);
169		return REG_EESCAPE;
170	}
171
172	{
173		char c = 0;
174		if (append(&buf, &c, 1, &size, &allocd) < 0)
175			goto fail;
176	}
177	*retp = buf;
178	return 0;
179}
180
181int
182globcomp(regex_t *preg, const char *glob, int cflags)
183{
184	char *regex = NULL;
185	int status = glob_to_regex(glob, &regex);
186	if (status != 0)
187		return status;
188	assert(regex != NULL);
189	status = regcomp(preg, regex, cflags);
190	free(regex);
191	return status;
192}
193
194#ifdef TEST
195#include <stdio.h>
196
197static void
198translate(const char *glob, int exp_status, const char *expect)
199{
200	char *pattern = NULL;
201	int status = glob_to_regex(glob, &pattern);
202	if (status != exp_status) {
203		fprintf(stderr, "translating %s, expected status %d, got %d\n",
204			glob, exp_status, status);
205		return;
206	}
207
208	if (status == 0) {
209		assert(pattern != NULL);
210		if (strcmp(pattern, expect) != 0)
211			fprintf(stderr, "translating %s, expected %s, got %s\n",
212				glob, expect, pattern);
213		free(pattern);
214	} else {
215		assert(pattern == NULL);
216	}
217}
218
219static void
220try_match(const char *glob, const char *str, int expect)
221{
222	regex_t preg;
223	int status = globcomp(&preg, glob, 0);
224	assert(status == 0);
225	status = regexec(&preg, str, 0, NULL, 0);
226	assert(status == expect);
227	regfree(&preg);
228}
229
230int
231main(void)
232{
233        translate("*", 0, ".*");
234        translate("?", 0, ".");
235        translate(".*", 0, "\\..*");
236        translate("*.*", 0, ".*\\..*");
237        translate("*a*", 0, ".*a.*");
238        translate("[abc]", 0, "[abc]");
239        translate("[^abc]", 0, "[^abc]");
240        translate("[!abc]", 0, "[^abc]");
241        translate("[]]", 0, "[]]");
242        translate("[[]", 0, "[[]");
243        translate("[^]]", 0, "[^]]");
244        translate("[^a-z]", 0, "[^a-z]");
245        translate("[abc\\]]", 0, "[abc\\]]");
246        translate("[abc\\]def]", 0, "[abc\\]def]");
247        translate("[[:space:]]", 0, "[[:space:]]");
248        translate("[^[:space:]]", 0, "[^[:space:]]");
249        translate("[![:space:]]", 0, "[^[:space:]]");
250        translate("[^a-z]*", 0, "[^a-z].*");
251        translate("[^a-z]bar*", 0, "[^a-z]bar.*");
252	translate("*.*.*.*.*.*.*.*.*.*.*.*.*.*.*.*.", 0,
253		  ".*\\..*\\..*\\..*\\..*\\..*\\..*\\..*\\."
254		  ".*\\..*\\..*\\..*\\..*\\..*\\..*\\..*\\.");
255
256        translate("\\", REG_EESCAPE, NULL);
257        translate("[^[:naotuh\\", REG_EBRACK, NULL);
258        translate("[^[:", REG_EBRACK, NULL);
259        translate("[^[", REG_EBRACK, NULL);
260        translate("[^", REG_EBRACK, NULL);
261        translate("[\\", REG_EBRACK, NULL);
262        translate("[", REG_EBRACK, NULL);
263        translate("abc[", REG_EBRACK, NULL);
264
265	try_match("abc*def", "abc012def", 0);
266	try_match("abc*def", "ab012def", REG_NOMATCH);
267	try_match("[abc]*def", "a1def", 0);
268	try_match("[abc]*def", "b1def", 0);
269	try_match("[abc]*def", "d1def", REG_NOMATCH);
270
271	return 0;
272}
273
274#endif
275