xref: /netbsd-src/external/bsd/libarchive/dist/libarchive/archive_string.c (revision 404fbe5fb94ca1e054339640cabb2801ce52dd30)
1 /*-
2  * Copyright (c) 2003-2007 Tim Kientzle
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(S) ``AS IS'' AND ANY EXPRESS OR
15  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
16  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
17  * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
18  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
19  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
20  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
21  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
22  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
23  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
24  */
25 
26 #include "archive_platform.h"
27 __FBSDID("$FreeBSD: src/lib/libarchive/archive_string.c,v 1.16 2008/06/15 11:28:56 kientzle Exp $");
28 
29 /*
30  * Basic resizable string support, to simplify manipulating arbitrary-sized
31  * strings while minimizing heap activity.
32  */
33 
34 #ifdef HAVE_STDLIB_H
35 #include <stdlib.h>
36 #endif
37 #ifdef HAVE_STRING_H
38 #include <string.h>
39 #endif
40 #ifdef HAVE_WCHAR_H
41 #include <wchar.h>
42 #endif
43 
44 #include "archive_private.h"
45 #include "archive_string.h"
46 
47 struct archive_string *
48 __archive_string_append(struct archive_string *as, const char *p, size_t s)
49 {
50 	if (__archive_string_ensure(as, as->length + s + 1) == NULL)
51 		__archive_errx(1, "Out of memory");
52 	memcpy(as->s + as->length, p, s);
53 	as->s[as->length + s] = 0;
54 	as->length += s;
55 	return (as);
56 }
57 
58 void
59 __archive_string_copy(struct archive_string *dest, struct archive_string *src)
60 {
61 	if (src->length == 0)
62 		dest->length = 0;
63 	else {
64 		if (__archive_string_ensure(dest, src->length + 1) == NULL)
65 			__archive_errx(1, "Out of memory");
66 		memcpy(dest->s, src->s, src->length);
67 		dest->length = src->length;
68 		dest->s[dest->length] = 0;
69 	}
70 }
71 
72 void
73 __archive_string_free(struct archive_string *as)
74 {
75 	as->length = 0;
76 	as->buffer_length = 0;
77 	if (as->s != NULL) {
78 		free(as->s);
79 		as->s = NULL;
80 	}
81 }
82 
83 /* Returns NULL on any allocation failure. */
84 struct archive_string *
85 __archive_string_ensure(struct archive_string *as, size_t s)
86 {
87 	/* If buffer is already big enough, don't reallocate. */
88 	if (as->s && (s <= as->buffer_length))
89 		return (as);
90 
91 	/*
92 	 * Growing the buffer at least exponentially ensures that
93 	 * append operations are always linear in the number of
94 	 * characters appended.  Using a smaller growth rate for
95 	 * larger buffers reduces memory waste somewhat at the cost of
96 	 * a larger constant factor.
97 	 */
98 	if (as->buffer_length < 32)
99 		/* Start with a minimum 32-character buffer. */
100 		as->buffer_length = 32;
101 	else if (as->buffer_length < 8192)
102 		/* Buffers under 8k are doubled for speed. */
103 		as->buffer_length *= 2;
104 	else {
105 		/* Buffers 8k and over grow by at least 25% each time. */
106 		size_t old_length = as->buffer_length;
107 		as->buffer_length = (as->buffer_length * 5) / 4;
108 		/* Be safe: If size wraps, release buffer and return NULL. */
109 		if (as->buffer_length < old_length) {
110 			free(as->s);
111 			as->s = NULL;
112 			return (NULL);
113 		}
114 	}
115 	/*
116 	 * The computation above is a lower limit to how much we'll
117 	 * grow the buffer.  In any case, we have to grow it enough to
118 	 * hold the request.
119 	 */
120 	if (as->buffer_length < s)
121 		as->buffer_length = s;
122 	/* Now we can reallocate the buffer. */
123 	as->s = (char *)realloc(as->s, as->buffer_length);
124 	if (as->s == NULL)
125 		return (NULL);
126 	return (as);
127 }
128 
129 struct archive_string *
130 __archive_strncat(struct archive_string *as, const char *p, size_t n)
131 {
132 	size_t s;
133 	const char *pp;
134 
135 	/* Like strlen(p), except won't examine positions beyond p[n]. */
136 	s = 0;
137 	pp = p;
138 	while (*pp && s < n) {
139 		pp++;
140 		s++;
141 	}
142 	return (__archive_string_append(as, p, s));
143 }
144 
145 struct archive_string *
146 __archive_strappend_char(struct archive_string *as, char c)
147 {
148 	return (__archive_string_append(as, &c, 1));
149 }
150 
151 struct archive_string *
152 __archive_strappend_int(struct archive_string *as, int d, int base)
153 {
154 	static const char *digits = "0123456789abcdef";
155 
156 	if (d < 0) {
157 		__archive_strappend_char(as, '-');
158 		d = -d;
159 	}
160 	if (d >= base)
161 		__archive_strappend_int(as, d/base, base);
162 	__archive_strappend_char(as, digits[d % base]);
163 	return (as);
164 }
165 
166 /*
167  * Home-grown wctomb for UTF-8.
168  */
169 static int
170 my_wctomb_utf8(char *p, wchar_t wc)
171 {
172 	if (p == NULL)
173 		/* UTF-8 doesn't use shift states. */
174 		return (0);
175 	if (wc <= 0x7f) {
176 		p[0] = (char)wc;
177 		return (1);
178 	}
179 	if (wc <= 0x7ff) {
180 		p[0] = 0xc0 | ((wc >> 6) & 0x1f);
181 		p[1] = 0x80 | (wc & 0x3f);
182 		return (2);
183 	}
184 	if (wc <= 0xffff) {
185 		p[0] = 0xe0 | ((wc >> 12) & 0x0f);
186 		p[1] = 0x80 | ((wc >> 6) & 0x3f);
187 		p[2] = 0x80 | (wc & 0x3f);
188 		return (3);
189 	}
190 	if (wc <= 0x1fffff) {
191 		p[0] = 0xf0 | ((wc >> 18) & 0x07);
192 		p[1] = 0x80 | ((wc >> 12) & 0x3f);
193 		p[2] = 0x80 | ((wc >> 6) & 0x3f);
194 		p[3] = 0x80 | (wc & 0x3f);
195 		return (4);
196 	}
197 	/* Unicode has no codes larger than 0x1fffff. */
198 	/*
199 	 * Awkward point:  UTF-8 <-> wchar_t conversions
200 	 * can actually fail.
201 	 */
202 	return (-1);
203 }
204 
205 static int
206 my_wcstombs(struct archive_string *as, const wchar_t *w,
207     int (*func)(char *, wchar_t))
208 {
209 	int n;
210 	char *p;
211 	char buff[256];
212 
213 	/* Clear the shift state before starting. */
214 	(*func)(NULL, L'\0');
215 
216 	/*
217 	 * Convert one wide char at a time into 'buff', whenever that
218 	 * fills, append it to the string.
219 	 */
220 	p = buff;
221 	while (*w != L'\0') {
222 		/* Flush the buffer when we have <=16 bytes free. */
223 		/* (No encoding has a single character >16 bytes.) */
224 		if ((size_t)(p - buff) >= (size_t)(sizeof(buff) - 16)) {
225 			*p = '\0';
226 			archive_strcat(as, buff);
227 			p = buff;
228 		}
229 		n = (*func)(p, *w++);
230 		if (n == -1)
231 			return (-1);
232 		p += n;
233 	}
234 	*p = '\0';
235 	archive_strcat(as, buff);
236 	return (0);
237 }
238 
239 /*
240  * Translates a wide character string into UTF-8 and appends
241  * to the archive_string.  Note: returns NULL if conversion fails.
242  */
243 struct archive_string *
244 __archive_strappend_w_utf8(struct archive_string *as, const wchar_t *w)
245 {
246 	if (my_wcstombs(as, w, my_wctomb_utf8))
247 		return (NULL);
248 	return (as);
249 }
250 
251 /*
252  * Translates a wide character string into current locale character set
253  * and appends to the archive_string.  Note: returns NULL if conversion
254  * fails.
255  */
256 struct archive_string *
257 __archive_strappend_w_mbs(struct archive_string *as, const wchar_t *w)
258 {
259 #if HAVE_WCTOMB
260 	if (my_wcstombs(as, w, wctomb))
261 		return (NULL);
262 #else
263 	/* TODO: Can we do better than this?  Are there platforms
264 	 * that have locale support but don't have wctomb()? */
265 	if (my_wcstombs(as, w, my_wctomb_utf8))
266 		return (NULL);
267 #endif
268 	return (as);
269 }
270 
271 
272 /*
273  * Home-grown mbtowc for UTF-8.  Some systems lack UTF-8
274  * (or even lack mbtowc()) and we need UTF-8 support for pax
275  * format.  So please don't replace this with a call to the
276  * standard mbtowc() function!
277  */
278 static int
279 my_mbtowc_utf8(wchar_t *pwc, const char *s, size_t n)
280 {
281         int ch;
282 
283 	/* Standard behavior:  a NULL value for 's' just resets shift state. */
284         if (s == NULL)
285                 return (0);
286 	/* If length argument is zero, don't look at the first character. */
287 	if (n <= 0)
288 		return (-1);
289 
290         /*
291 	 * Decode 1-4 bytes depending on the value of the first byte.
292 	 */
293         ch = (unsigned char)*s;
294 	if (ch == 0) {
295 		return (0); /* Standard:  return 0 for end-of-string. */
296 	}
297 	if ((ch & 0x80) == 0) {
298                 *pwc = ch & 0x7f;
299 		return (1);
300         }
301 	if ((ch & 0xe0) == 0xc0) {
302 		if (n < 2)
303 			return (-1);
304 		if ((s[1] & 0xc0) != 0x80) return (-1);
305                 *pwc = ((ch & 0x1f) << 6) | (s[1] & 0x3f);
306 		return (2);
307         }
308 	if ((ch & 0xf0) == 0xe0) {
309 		if (n < 3)
310 			return (-1);
311 		if ((s[1] & 0xc0) != 0x80) return (-1);
312 		if ((s[2] & 0xc0) != 0x80) return (-1);
313                 *pwc = ((ch & 0x0f) << 12)
314 		    | ((s[1] & 0x3f) << 6)
315 		    | (s[2] & 0x3f);
316 		return (3);
317         }
318 	if ((ch & 0xf8) == 0xf0) {
319 		if (n < 4)
320 			return (-1);
321 		if ((s[1] & 0xc0) != 0x80) return (-1);
322 		if ((s[2] & 0xc0) != 0x80) return (-1);
323 		if ((s[3] & 0xc0) != 0x80) return (-1);
324                 *pwc = ((ch & 0x07) << 18)
325 		    | ((s[1] & 0x3f) << 12)
326 		    | ((s[2] & 0x3f) << 6)
327 		    | (s[3] & 0x3f);
328 		return (4);
329         }
330 	/* Invalid first byte. */
331 	return (-1);
332 }
333 
334 /*
335  * Return a wide-character string by converting this archive_string
336  * from UTF-8.
337  */
338 wchar_t *
339 __archive_string_utf8_w(struct archive_string *as)
340 {
341 	wchar_t *ws, *dest;
342 	const char *src;
343 	int n;
344 	int err;
345 
346 	ws = (wchar_t *)malloc((as->length + 1) * sizeof(wchar_t));
347 	if (ws == NULL)
348 		__archive_errx(1, "Out of memory");
349 	err = 0;
350 	dest = ws;
351 	src = as->s;
352 	while (*src != '\0') {
353 		n = my_mbtowc_utf8(dest, src, 8);
354 		if (n == 0)
355 			break;
356 		if (n < 0) {
357 			free(ws);
358 			return (NULL);
359 		}
360 		dest++;
361 		src += n;
362 	}
363 	*dest++ = L'\0';
364 	return (ws);
365 }
366