xref: /freebsd/contrib/bmake/str.c (revision e8d8bef961a50d4dc22501cde4fb9fb0be1b2532)
1 /*	$NetBSD: str.c,v 1.81 2021/02/01 22:36:28 rillig Exp $	*/
2 
3 /*
4  * Copyright (c) 1988, 1989, 1990, 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 de Boor.
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 /*
36  * Copyright (c) 1989 by Berkeley Softworks
37  * All rights reserved.
38  *
39  * This code is derived from software contributed to Berkeley by
40  * Adam de Boor.
41  *
42  * Redistribution and use in source and binary forms, with or without
43  * modification, are permitted provided that the following conditions
44  * are met:
45  * 1. Redistributions of source code must retain the above copyright
46  *    notice, this list of conditions and the following disclaimer.
47  * 2. Redistributions in binary form must reproduce the above copyright
48  *    notice, this list of conditions and the following disclaimer in the
49  *    documentation and/or other materials provided with the distribution.
50  * 3. All advertising materials mentioning features or use of this software
51  *    must display the following acknowledgement:
52  *	This product includes software developed by the University of
53  *	California, Berkeley and its contributors.
54  * 4. Neither the name of the University nor the names of its contributors
55  *    may be used to endorse or promote products derived from this software
56  *    without specific prior written permission.
57  *
58  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
59  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
60  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
61  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
62  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
63  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
64  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
65  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
66  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
67  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
68  * SUCH DAMAGE.
69  */
70 
71 #include "make.h"
72 
73 /*	"@(#)str.c	5.8 (Berkeley) 6/1/90"	*/
74 MAKE_RCSID("$NetBSD: str.c,v 1.81 2021/02/01 22:36:28 rillig Exp $");
75 
76 /* Return the concatenation of s1 and s2, freshly allocated. */
77 char *
78 str_concat2(const char *s1, const char *s2)
79 {
80 	size_t len1 = strlen(s1);
81 	size_t len2 = strlen(s2);
82 	char *result = bmake_malloc(len1 + len2 + 1);
83 	memcpy(result, s1, len1);
84 	memcpy(result + len1, s2, len2 + 1);
85 	return result;
86 }
87 
88 /* Return the concatenation of s1, s2 and s3, freshly allocated. */
89 char *
90 str_concat3(const char *s1, const char *s2, const char *s3)
91 {
92 	size_t len1 = strlen(s1);
93 	size_t len2 = strlen(s2);
94 	size_t len3 = strlen(s3);
95 	char *result = bmake_malloc(len1 + len2 + len3 + 1);
96 	memcpy(result, s1, len1);
97 	memcpy(result + len1, s2, len2);
98 	memcpy(result + len1 + len2, s3, len3 + 1);
99 	return result;
100 }
101 
102 /* Return the concatenation of s1, s2, s3 and s4, freshly allocated. */
103 char *
104 str_concat4(const char *s1, const char *s2, const char *s3, const char *s4)
105 {
106 	size_t len1 = strlen(s1);
107 	size_t len2 = strlen(s2);
108 	size_t len3 = strlen(s3);
109 	size_t len4 = strlen(s4);
110 	char *result = bmake_malloc(len1 + len2 + len3 + len4 + 1);
111 	memcpy(result, s1, len1);
112 	memcpy(result + len1, s2, len2);
113 	memcpy(result + len1 + len2, s3, len3);
114 	memcpy(result + len1 + len2 + len3, s4, len4 + 1);
115 	return result;
116 }
117 
118 /*
119  * Fracture a string into an array of words (as delineated by tabs or spaces)
120  * taking quotation marks into account.
121  *
122  * If expand is TRUE, quotes are removed and escape sequences such as \r, \t,
123  * etc... are expanded. In this case, return NULL on parse errors.
124  *
125  * Returns the fractured words, which must be freed later using Words_Free,
126  * unless the returned Words.words was NULL.
127  */
128 Words
129 Str_Words(const char *str, Boolean expand)
130 {
131 	size_t str_len;
132 	char *words_buf;
133 	size_t words_cap;
134 	char **words;
135 	size_t words_len;
136 	char inquote;
137 	char *word_start;
138 	char *word_end;
139 	const char *str_p;
140 
141 	/* XXX: why only hspace, not whitespace? */
142 	cpp_skip_hspace(&str);	/* skip leading space chars. */
143 
144 	/* words_buf holds the words, separated by '\0'. */
145 	str_len = strlen(str);
146 	words_buf = bmake_malloc(str_len + 1);
147 
148 	words_cap = str_len / 5 > 50 ? str_len / 5 : 50;
149 	words = bmake_malloc((words_cap + 1) * sizeof(char *));
150 
151 	/*
152 	 * copy the string; at the same time, parse backslashes,
153 	 * quotes and build the word list.
154 	 */
155 	words_len = 0;
156 	inquote = '\0';
157 	word_start = words_buf;
158 	word_end = words_buf;
159 	for (str_p = str;; str_p++) {
160 		char ch = *str_p;
161 		switch (ch) {
162 		case '"':
163 		case '\'':
164 			if (inquote != '\0') {
165 				if (inquote == ch)
166 					inquote = '\0';
167 				else
168 					break;
169 			} else {
170 				inquote = ch;
171 				/* Don't miss "" or '' */
172 				if (word_start == NULL && str_p[1] == inquote) {
173 					if (!expand) {
174 						word_start = word_end;
175 						*word_end++ = ch;
176 					} else
177 						word_start = word_end + 1;
178 					str_p++;
179 					inquote = '\0';
180 					break;
181 				}
182 			}
183 			if (!expand) {
184 				if (word_start == NULL)
185 					word_start = word_end;
186 				*word_end++ = ch;
187 			}
188 			continue;
189 		case ' ':
190 		case '\t':
191 		case '\n':
192 			if (inquote != '\0')
193 				break;
194 			if (word_start == NULL)
195 				continue;
196 			/* FALLTHROUGH */
197 		case '\0':
198 			/*
199 			 * end of a token -- make sure there's enough words
200 			 * space and save off a pointer.
201 			 */
202 			if (word_start == NULL)
203 				goto done;
204 
205 			*word_end++ = '\0';
206 			if (words_len == words_cap) {
207 				size_t new_size;
208 				words_cap *= 2;		/* ramp up fast */
209 				new_size = (words_cap + 1) * sizeof(char *);
210 				words = bmake_realloc(words, new_size);
211 			}
212 			words[words_len++] = word_start;
213 			word_start = NULL;
214 			if (ch == '\n' || ch == '\0') {
215 				if (expand && inquote != '\0') {
216 					free(words);
217 					free(words_buf);
218 					return (Words){ NULL, 0, NULL };
219 				}
220 				goto done;
221 			}
222 			continue;
223 		case '\\':
224 			if (!expand) {
225 				if (word_start == NULL)
226 					word_start = word_end;
227 				*word_end++ = '\\';
228 				/* catch '\' at end of line */
229 				if (str_p[1] == '\0')
230 					continue;
231 				ch = *++str_p;
232 				break;
233 			}
234 
235 			switch (ch = *++str_p) {
236 			case '\0':
237 			case '\n':
238 				/* hmmm; fix it up as best we can */
239 				ch = '\\';
240 				str_p--;
241 				break;
242 			case 'b':
243 				ch = '\b';
244 				break;
245 			case 'f':
246 				ch = '\f';
247 				break;
248 			case 'n':
249 				ch = '\n';
250 				break;
251 			case 'r':
252 				ch = '\r';
253 				break;
254 			case 't':
255 				ch = '\t';
256 				break;
257 			}
258 			break;
259 		}
260 		if (word_start == NULL)
261 			word_start = word_end;
262 		*word_end++ = ch;
263 	}
264 done:
265 	words[words_len] = NULL;	/* useful for argv */
266 	return (Words){ words, words_len, words_buf };
267 }
268 
269 /*
270  * Str_Match -- Test if a string matches a pattern like "*.[ch]".
271  * The following special characters are known *?\[] (as in fnmatch(3)).
272  *
273  * XXX: this function does not detect or report malformed patterns.
274  */
275 Boolean
276 Str_Match(const char *str, const char *pat)
277 {
278 	for (;;) {
279 		/*
280 		 * See if we're at the end of both the pattern and the
281 		 * string. If so, we succeeded.  If we're at the end of the
282 		 * pattern but not at the end of the string, we failed.
283 		 */
284 		if (*pat == '\0')
285 			return *str == '\0';
286 		if (*str == '\0' && *pat != '*')
287 			return FALSE;
288 
289 		/*
290 		 * A '*' in the pattern matches any substring.  We handle this
291 		 * by calling ourselves for each suffix of the string.
292 		 */
293 		if (*pat == '*') {
294 			pat++;
295 			while (*pat == '*')
296 				pat++;
297 			if (*pat == '\0')
298 				return TRUE;
299 			while (*str != '\0') {
300 				if (Str_Match(str, pat))
301 					return TRUE;
302 				str++;
303 			}
304 			return FALSE;
305 		}
306 
307 		/* A '?' in the pattern matches any single character. */
308 		if (*pat == '?')
309 			goto thisCharOK;
310 
311 		/*
312 		 * A '[' in the pattern matches a character from a list.
313 		 * The '[' is followed by the list of acceptable characters,
314 		 * or by ranges (two characters separated by '-'). In these
315 		 * character lists, the backslash is an ordinary character.
316 		 */
317 		if (*pat == '[') {
318 			Boolean neg = pat[1] == '^';
319 			pat += neg ? 2 : 1;
320 
321 			for (;;) {
322 				if (*pat == ']' || *pat == '\0') {
323 					if (neg)
324 						break;
325 					return FALSE;
326 				}
327 				/*
328 				 * XXX: This naive comparison makes the
329 				 * control flow of the pattern parser
330 				 * dependent on the actual value of the
331 				 * string.  This is unpredictable.  It may be
332 				 * though that the code only looks wrong but
333 				 * actually all code paths result in the same
334 				 * behavior.  This needs further tests.
335 				 */
336 				if (*pat == *str)
337 					break;
338 				if (pat[1] == '-') {
339 					if (pat[2] == '\0')
340 						return neg;
341 					if (*pat <= *str && pat[2] >= *str)
342 						break;
343 					if (*pat >= *str && pat[2] <= *str)
344 						break;
345 					pat += 2;
346 				}
347 				pat++;
348 			}
349 			if (neg && *pat != ']' && *pat != '\0')
350 				return FALSE;
351 			while (*pat != ']' && *pat != '\0')
352 				pat++;
353 			if (*pat == '\0')
354 				pat--;
355 			goto thisCharOK;
356 		}
357 
358 		/*
359 		 * A backslash in the pattern matches the character following
360 		 * it exactly.
361 		 */
362 		if (*pat == '\\') {
363 			pat++;
364 			if (*pat == '\0')
365 				return FALSE;
366 		}
367 
368 		if (*pat != *str)
369 			return FALSE;
370 
371 	thisCharOK:
372 		pat++;
373 		str++;
374 	}
375 }
376