unzip.c revision 1.25 1 1.25 christos /* $NetBSD: unzip.c,v 1.25 2021/02/18 17:05:51 christos Exp $ */
2 1.1 joerg
3 1.1 joerg /*-
4 1.13 joerg * Copyright (c) 2009, 2010 Joerg Sonnenberger <joerg (at) NetBSD.org>
5 1.1 joerg * Copyright (c) 2007-2008 Dag-Erling Codan Smrgrav
6 1.1 joerg * All rights reserved.
7 1.1 joerg *
8 1.1 joerg * Redistribution and use in source and binary forms, with or without
9 1.1 joerg * modification, are permitted provided that the following conditions
10 1.1 joerg * are met:
11 1.1 joerg * 1. Redistributions of source code must retain the above copyright
12 1.1 joerg * notice, this list of conditions and the following disclaimer
13 1.1 joerg * in this position and unchanged.
14 1.1 joerg * 2. Redistributions in binary form must reproduce the above copyright
15 1.1 joerg * notice, this list of conditions and the following disclaimer in the
16 1.1 joerg * documentation and/or other materials provided with the distribution.
17 1.1 joerg *
18 1.1 joerg * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19 1.1 joerg * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 1.1 joerg * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 1.1 joerg * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
22 1.1 joerg * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23 1.1 joerg * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24 1.1 joerg * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25 1.1 joerg * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 1.1 joerg * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27 1.1 joerg * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 1.1 joerg * SUCH DAMAGE.
29 1.1 joerg *
30 1.1 joerg * $FreeBSD: revision 180124$
31 1.1 joerg *
32 1.1 joerg * This file would be much shorter if we didn't care about command-line
33 1.1 joerg * compatibility with Info-ZIP's UnZip, which requires us to duplicate
34 1.1 joerg * parts of libarchive in order to gain more detailed control of its
35 1.1 joerg * behaviour for the purpose of implementing the -n, -o, -L and -a
36 1.1 joerg * options.
37 1.1 joerg */
38 1.1 joerg
39 1.1 joerg #include <sys/cdefs.h>
40 1.25 christos __RCSID("$NetBSD: unzip.c,v 1.25 2021/02/18 17:05:51 christos Exp $");
41 1.25 christos
42 1.25 christos #ifdef __GLIBC__
43 1.25 christos #define _GNU_SOURCE
44 1.25 christos #endif
45 1.1 joerg
46 1.1 joerg #include <sys/queue.h>
47 1.1 joerg #include <sys/stat.h>
48 1.1 joerg
49 1.1 joerg #include <ctype.h>
50 1.1 joerg #include <errno.h>
51 1.1 joerg #include <fcntl.h>
52 1.1 joerg #include <fnmatch.h>
53 1.1 joerg #include <stdarg.h>
54 1.1 joerg #include <stdio.h>
55 1.1 joerg #include <stdlib.h>
56 1.1 joerg #include <string.h>
57 1.1 joerg #include <unistd.h>
58 1.1 joerg
59 1.1 joerg #include <archive.h>
60 1.1 joerg #include <archive_entry.h>
61 1.1 joerg
62 1.1 joerg /* command-line options */
63 1.1 joerg static int a_opt; /* convert EOL */
64 1.7 wiz static int C_opt; /* match case-insensitively */
65 1.4 wiz static int c_opt; /* extract to stdout */
66 1.1 joerg static const char *d_arg; /* directory */
67 1.1 joerg static int f_opt; /* update existing files only */
68 1.1 joerg static int j_opt; /* junk directories */
69 1.1 joerg static int L_opt; /* lowercase names */
70 1.1 joerg static int n_opt; /* never overwrite */
71 1.1 joerg static int o_opt; /* always overwrite */
72 1.3 joerg static int p_opt; /* extract to stdout, quiet */
73 1.1 joerg static int q_opt; /* quiet */
74 1.1 joerg static int t_opt; /* test */
75 1.1 joerg static int u_opt; /* update */
76 1.3 joerg static int v_opt; /* verbose/list */
77 1.17 christos static const char * y_str = ""; /* 4 digit year */
78 1.1 joerg
79 1.1 joerg /* time when unzip started */
80 1.1 joerg static time_t now;
81 1.1 joerg
82 1.1 joerg /* debug flag */
83 1.1 joerg static int unzip_debug;
84 1.1 joerg
85 1.1 joerg /* running on tty? */
86 1.1 joerg static int tty;
87 1.1 joerg
88 1.1 joerg /* convenience macro */
89 1.1 joerg /* XXX should differentiate between ARCHIVE_{WARN,FAIL,RETRY} */
90 1.1 joerg #define ac(call) \
91 1.1 joerg do { \
92 1.1 joerg int acret = (call); \
93 1.1 joerg if (acret != ARCHIVE_OK) \
94 1.1 joerg errorx("%s", archive_error_string(a)); \
95 1.1 joerg } while (0)
96 1.1 joerg
97 1.1 joerg /*
98 1.1 joerg * Indicates that last info() did not end with EOL. This helps error() et
99 1.1 joerg * al. avoid printing an error message on the same line as an incomplete
100 1.1 joerg * informational message.
101 1.1 joerg */
102 1.1 joerg static int noeol;
103 1.1 joerg
104 1.1 joerg /* fatal error message + errno */
105 1.18 joerg __dead __printflike(1, 2) static void
106 1.1 joerg error(const char *fmt, ...)
107 1.1 joerg {
108 1.1 joerg va_list ap;
109 1.1 joerg
110 1.1 joerg if (noeol)
111 1.1 joerg fprintf(stdout, "\n");
112 1.1 joerg fflush(stdout);
113 1.1 joerg fprintf(stderr, "unzip: ");
114 1.1 joerg va_start(ap, fmt);
115 1.1 joerg vfprintf(stderr, fmt, ap);
116 1.1 joerg va_end(ap);
117 1.1 joerg fprintf(stderr, ": %s\n", strerror(errno));
118 1.1 joerg exit(1);
119 1.1 joerg }
120 1.1 joerg
121 1.1 joerg /* fatal error message, no errno */
122 1.18 joerg __dead __printflike(1, 2) static void
123 1.1 joerg errorx(const char *fmt, ...)
124 1.1 joerg {
125 1.1 joerg va_list ap;
126 1.1 joerg
127 1.1 joerg if (noeol)
128 1.1 joerg fprintf(stdout, "\n");
129 1.1 joerg fflush(stdout);
130 1.1 joerg fprintf(stderr, "unzip: ");
131 1.1 joerg va_start(ap, fmt);
132 1.1 joerg vfprintf(stderr, fmt, ap);
133 1.1 joerg va_end(ap);
134 1.1 joerg fprintf(stderr, "\n");
135 1.1 joerg exit(1);
136 1.1 joerg }
137 1.1 joerg
138 1.1 joerg /* non-fatal error message + errno */
139 1.18 joerg __printflike(1, 2) static void
140 1.1 joerg warning(const char *fmt, ...)
141 1.1 joerg {
142 1.1 joerg va_list ap;
143 1.1 joerg
144 1.1 joerg if (noeol)
145 1.1 joerg fprintf(stdout, "\n");
146 1.1 joerg fflush(stdout);
147 1.1 joerg fprintf(stderr, "unzip: ");
148 1.1 joerg va_start(ap, fmt);
149 1.1 joerg vfprintf(stderr, fmt, ap);
150 1.1 joerg va_end(ap);
151 1.1 joerg fprintf(stderr, ": %s\n", strerror(errno));
152 1.1 joerg }
153 1.20 christos
154 1.1 joerg /* non-fatal error message, no errno */
155 1.18 joerg __printflike(1, 2) static void
156 1.1 joerg warningx(const char *fmt, ...)
157 1.1 joerg {
158 1.1 joerg va_list ap;
159 1.1 joerg
160 1.1 joerg if (noeol)
161 1.1 joerg fprintf(stdout, "\n");
162 1.1 joerg fflush(stdout);
163 1.1 joerg fprintf(stderr, "unzip: ");
164 1.1 joerg va_start(ap, fmt);
165 1.1 joerg vfprintf(stderr, fmt, ap);
166 1.1 joerg va_end(ap);
167 1.1 joerg fprintf(stderr, "\n");
168 1.1 joerg }
169 1.1 joerg
170 1.1 joerg /* informational message (if not -q) */
171 1.18 joerg __printflike(1, 2) static void
172 1.1 joerg info(const char *fmt, ...)
173 1.1 joerg {
174 1.1 joerg va_list ap;
175 1.1 joerg
176 1.1 joerg if (q_opt && !unzip_debug)
177 1.1 joerg return;
178 1.1 joerg va_start(ap, fmt);
179 1.1 joerg vfprintf(stdout, fmt, ap);
180 1.1 joerg va_end(ap);
181 1.1 joerg fflush(stdout);
182 1.1 joerg
183 1.1 joerg if (*fmt == '\0')
184 1.1 joerg noeol = 1;
185 1.1 joerg else
186 1.1 joerg noeol = fmt[strlen(fmt) - 1] != '\n';
187 1.1 joerg }
188 1.1 joerg
189 1.1 joerg /* debug message (if unzip_debug) */
190 1.18 joerg __printflike(1, 2) static void
191 1.1 joerg debug(const char *fmt, ...)
192 1.1 joerg {
193 1.1 joerg va_list ap;
194 1.1 joerg
195 1.1 joerg if (!unzip_debug)
196 1.1 joerg return;
197 1.1 joerg va_start(ap, fmt);
198 1.1 joerg vfprintf(stderr, fmt, ap);
199 1.1 joerg va_end(ap);
200 1.1 joerg fflush(stderr);
201 1.1 joerg
202 1.1 joerg if (*fmt == '\0')
203 1.1 joerg noeol = 1;
204 1.1 joerg else
205 1.1 joerg noeol = fmt[strlen(fmt) - 1] != '\n';
206 1.1 joerg }
207 1.1 joerg
208 1.1 joerg /* duplicate a path name, possibly converting to lower case */
209 1.1 joerg static char *
210 1.1 joerg pathdup(const char *path)
211 1.1 joerg {
212 1.1 joerg char *str;
213 1.1 joerg size_t i, len;
214 1.1 joerg
215 1.1 joerg len = strlen(path);
216 1.1 joerg while (len && path[len - 1] == '/')
217 1.1 joerg len--;
218 1.1 joerg if ((str = malloc(len + 1)) == NULL) {
219 1.1 joerg errno = ENOMEM;
220 1.1 joerg error("malloc()");
221 1.1 joerg }
222 1.1 joerg if (L_opt) {
223 1.1 joerg for (i = 0; i < len; ++i)
224 1.1 joerg str[i] = tolower((unsigned char)path[i]);
225 1.1 joerg } else {
226 1.1 joerg memcpy(str, path, len);
227 1.1 joerg }
228 1.1 joerg str[len] = '\0';
229 1.1 joerg
230 1.1 joerg return (str);
231 1.1 joerg }
232 1.1 joerg
233 1.1 joerg /* concatenate two path names */
234 1.1 joerg static char *
235 1.1 joerg pathcat(const char *prefix, const char *path)
236 1.1 joerg {
237 1.1 joerg char *str;
238 1.1 joerg size_t prelen, len;
239 1.1 joerg
240 1.1 joerg prelen = prefix ? strlen(prefix) + 1 : 0;
241 1.1 joerg len = strlen(path) + 1;
242 1.1 joerg if ((str = malloc(prelen + len)) == NULL) {
243 1.1 joerg errno = ENOMEM;
244 1.1 joerg error("malloc()");
245 1.1 joerg }
246 1.1 joerg if (prefix) {
247 1.1 joerg memcpy(str, prefix, prelen); /* includes zero */
248 1.1 joerg str[prelen - 1] = '/'; /* splat zero */
249 1.1 joerg }
250 1.1 joerg memcpy(str + prelen, path, len); /* includes zero */
251 1.1 joerg
252 1.1 joerg return (str);
253 1.1 joerg }
254 1.1 joerg
255 1.1 joerg /*
256 1.1 joerg * Pattern lists for include / exclude processing
257 1.1 joerg */
258 1.1 joerg struct pattern {
259 1.1 joerg STAILQ_ENTRY(pattern) link;
260 1.1 joerg char pattern[];
261 1.1 joerg };
262 1.1 joerg
263 1.1 joerg STAILQ_HEAD(pattern_list, pattern);
264 1.1 joerg static struct pattern_list include = STAILQ_HEAD_INITIALIZER(include);
265 1.1 joerg static struct pattern_list exclude = STAILQ_HEAD_INITIALIZER(exclude);
266 1.1 joerg
267 1.1 joerg /*
268 1.1 joerg * Add an entry to a pattern list
269 1.1 joerg */
270 1.1 joerg static void
271 1.1 joerg add_pattern(struct pattern_list *list, const char *pattern)
272 1.1 joerg {
273 1.1 joerg struct pattern *entry;
274 1.1 joerg size_t len;
275 1.1 joerg
276 1.1 joerg debug("adding pattern '%s'\n", pattern);
277 1.1 joerg len = strlen(pattern);
278 1.1 joerg if ((entry = malloc(sizeof *entry + len + 1)) == NULL) {
279 1.1 joerg errno = ENOMEM;
280 1.1 joerg error("malloc()");
281 1.1 joerg }
282 1.1 joerg memcpy(entry->pattern, pattern, len + 1);
283 1.1 joerg STAILQ_INSERT_TAIL(list, entry, link);
284 1.1 joerg }
285 1.1 joerg
286 1.1 joerg /*
287 1.1 joerg * Match a string against a list of patterns
288 1.1 joerg */
289 1.1 joerg static int
290 1.1 joerg match_pattern(struct pattern_list *list, const char *str)
291 1.1 joerg {
292 1.1 joerg struct pattern *entry;
293 1.1 joerg
294 1.1 joerg STAILQ_FOREACH(entry, list, link) {
295 1.7 wiz if (fnmatch(entry->pattern, str, C_opt ? FNM_CASEFOLD : 0) == 0)
296 1.1 joerg return (1);
297 1.1 joerg }
298 1.1 joerg return (0);
299 1.1 joerg }
300 1.1 joerg
301 1.1 joerg /*
302 1.1 joerg * Verify that a given pathname is in the include list and not in the
303 1.1 joerg * exclude list.
304 1.1 joerg */
305 1.1 joerg static int
306 1.1 joerg accept_pathname(const char *pathname)
307 1.1 joerg {
308 1.1 joerg
309 1.1 joerg if (!STAILQ_EMPTY(&include) && !match_pattern(&include, pathname))
310 1.1 joerg return (0);
311 1.1 joerg if (!STAILQ_EMPTY(&exclude) && match_pattern(&exclude, pathname))
312 1.1 joerg return (0);
313 1.1 joerg return (1);
314 1.1 joerg }
315 1.1 joerg
316 1.1 joerg /*
317 1.1 joerg * Create the specified directory with the specified mode, taking certain
318 1.1 joerg * precautions on they way.
319 1.1 joerg */
320 1.1 joerg static void
321 1.1 joerg make_dir(const char *path, int mode)
322 1.1 joerg {
323 1.1 joerg struct stat sb;
324 1.1 joerg
325 1.1 joerg if (lstat(path, &sb) == 0) {
326 1.1 joerg if (S_ISDIR(sb.st_mode))
327 1.1 joerg return;
328 1.1 joerg /*
329 1.1 joerg * Normally, we should either ask the user about removing
330 1.1 joerg * the non-directory of the same name as a directory we
331 1.1 joerg * wish to create, or respect the -n or -o command-line
332 1.1 joerg * options. However, this may lead to a later failure or
333 1.1 joerg * even compromise (if this non-directory happens to be a
334 1.1 joerg * symlink to somewhere unsafe), so we don't.
335 1.1 joerg */
336 1.1 joerg
337 1.1 joerg /*
338 1.1 joerg * Don't check unlink() result; failure will cause mkdir()
339 1.1 joerg * to fail later, which we will catch.
340 1.1 joerg */
341 1.1 joerg (void)unlink(path);
342 1.1 joerg }
343 1.1 joerg if (mkdir(path, mode) != 0 && errno != EEXIST)
344 1.1 joerg error("mkdir('%s')", path);
345 1.1 joerg }
346 1.1 joerg
347 1.1 joerg /*
348 1.1 joerg * Ensure that all directories leading up to (but not including) the
349 1.1 joerg * specified path exist.
350 1.1 joerg *
351 1.1 joerg * XXX inefficient + modifies the file in-place
352 1.1 joerg */
353 1.1 joerg static void
354 1.1 joerg make_parent(char *path)
355 1.1 joerg {
356 1.1 joerg struct stat sb;
357 1.1 joerg char *sep;
358 1.1 joerg
359 1.1 joerg sep = strrchr(path, '/');
360 1.1 joerg if (sep == NULL || sep == path)
361 1.1 joerg return;
362 1.1 joerg *sep = '\0';
363 1.1 joerg if (lstat(path, &sb) == 0) {
364 1.1 joerg if (S_ISDIR(sb.st_mode)) {
365 1.1 joerg *sep = '/';
366 1.1 joerg return;
367 1.1 joerg }
368 1.1 joerg unlink(path);
369 1.1 joerg }
370 1.1 joerg make_parent(path);
371 1.1 joerg mkdir(path, 0755);
372 1.1 joerg *sep = '/';
373 1.1 joerg
374 1.1 joerg #if 0
375 1.1 joerg for (sep = path; (sep = strchr(sep, '/')) != NULL; sep++) {
376 1.1 joerg /* root in case of absolute d_arg */
377 1.1 joerg if (sep == path)
378 1.1 joerg continue;
379 1.1 joerg *sep = '\0';
380 1.1 joerg make_dir(path, 0755);
381 1.1 joerg *sep = '/';
382 1.1 joerg }
383 1.1 joerg #endif
384 1.1 joerg }
385 1.1 joerg
386 1.1 joerg /*
387 1.1 joerg * Extract a directory.
388 1.1 joerg */
389 1.1 joerg static void
390 1.1 joerg extract_dir(struct archive *a, struct archive_entry *e, const char *path)
391 1.1 joerg {
392 1.1 joerg int mode;
393 1.1 joerg
394 1.24 joerg /*
395 1.24 joerg * Dropbox likes to create '/' directory entries, just ignore
396 1.24 joerg * such junk.
397 1.24 joerg */
398 1.24 joerg if (*path == '\0')
399 1.24 joerg return;
400 1.24 joerg
401 1.11 mbalmer mode = archive_entry_mode(e) & 0777;
402 1.1 joerg if (mode == 0)
403 1.1 joerg mode = 0755;
404 1.1 joerg
405 1.1 joerg /*
406 1.1 joerg * Some zipfiles contain directories with weird permissions such
407 1.1 joerg * as 0644 or 0444. This can cause strange issues such as being
408 1.1 joerg * unable to extract files into the directory we just created, or
409 1.1 joerg * the user being unable to remove the directory later without
410 1.1 joerg * first manually changing its permissions. Therefore, we whack
411 1.1 joerg * the permissions into shape, assuming that the user wants full
412 1.1 joerg * access and that anyone who gets read access also gets execute
413 1.1 joerg * access.
414 1.1 joerg */
415 1.1 joerg mode |= 0700;
416 1.1 joerg if (mode & 0040)
417 1.1 joerg mode |= 0010;
418 1.1 joerg if (mode & 0004)
419 1.1 joerg mode |= 0001;
420 1.1 joerg
421 1.10 wiz info(" creating: %s/\n", path);
422 1.1 joerg make_dir(path, mode);
423 1.1 joerg ac(archive_read_data_skip(a));
424 1.1 joerg }
425 1.1 joerg
426 1.1 joerg static unsigned char buffer[8192];
427 1.1 joerg static char spinner[] = { '|', '/', '-', '\\' };
428 1.1 joerg
429 1.8 joerg static int
430 1.8 joerg handle_existing_file(char **path)
431 1.8 joerg {
432 1.8 joerg size_t alen;
433 1.8 joerg ssize_t len;
434 1.8 joerg char buf[4];
435 1.8 joerg
436 1.8 joerg for (;;) {
437 1.8 joerg fprintf(stderr,
438 1.8 joerg "replace %s? [y]es, [n]o, [A]ll, [N]one, [r]ename: ",
439 1.8 joerg *path);
440 1.8 joerg fgets(buf, 4, stdin);
441 1.8 joerg switch (*buf) {
442 1.8 joerg case 'A':
443 1.8 joerg o_opt = 1;
444 1.8 joerg /* FALL THROUGH */
445 1.8 joerg case 'y':
446 1.8 joerg case 'Y':
447 1.8 joerg (void)unlink(*path);
448 1.8 joerg return 1;
449 1.8 joerg case 'N':
450 1.22 christos n_opt = 1;
451 1.8 joerg /* FALL THROUGH */
452 1.8 joerg case 'n':
453 1.8 joerg return -1;
454 1.8 joerg case 'r':
455 1.8 joerg case 'R':
456 1.8 joerg printf("New name: ");
457 1.8 joerg fflush(stdout);
458 1.8 joerg free(*path);
459 1.8 joerg *path = NULL;
460 1.8 joerg alen = 0;
461 1.8 joerg len = getline(path, &alen, stdin);
462 1.15 wiz if ((*path)[len - 1] == '\n')
463 1.8 joerg (*path)[len - 1] = '\0';
464 1.8 joerg return 0;
465 1.8 joerg default:
466 1.8 joerg break;
467 1.8 joerg }
468 1.8 joerg }
469 1.8 joerg }
470 1.8 joerg
471 1.1 joerg /*
472 1.13 joerg * Detect binary files by a combination of character white list and
473 1.13 joerg * black list. NUL bytes and other control codes without use in text files
474 1.13 joerg * result directly in switching the file to binary mode. Otherwise, at least
475 1.13 joerg * one white-listed byte has to be found.
476 1.13 joerg *
477 1.13 joerg * Black-listed: 0..6, 14..25, 28..31
478 1.13 joerg * White-listed: 9..10, 13, >= 32
479 1.13 joerg *
480 1.13 joerg * See the proginfo/txtvsbin.txt in the zip sources for a detailed discussion.
481 1.13 joerg */
482 1.13 joerg #define BYTE_IS_BINARY(x) ((x) < 32 && (0xf3ffc07fU & (1U << (x))))
483 1.13 joerg #define BYTE_IS_TEXT(x) ((x) >= 32 || (0x00002600U & (1U << (x))))
484 1.13 joerg
485 1.13 joerg static int
486 1.13 joerg check_binary(const unsigned char *buf, size_t len)
487 1.13 joerg {
488 1.13 joerg int rv;
489 1.13 joerg for (rv = 1; len--; ++buf) {
490 1.13 joerg if (BYTE_IS_BINARY(*buf))
491 1.13 joerg return 1;
492 1.13 joerg if (BYTE_IS_TEXT(*buf))
493 1.13 joerg rv = 0;
494 1.13 joerg }
495 1.13 joerg
496 1.13 joerg return rv;
497 1.13 joerg }
498 1.13 joerg
499 1.13 joerg /*
500 1.22 christos * Extract to a file descriptor
501 1.22 christos */
502 1.22 christos static int
503 1.22 christos extract2fd(struct archive *a, char *pathname, int fd)
504 1.22 christos {
505 1.22 christos int cr, text, warn;
506 1.22 christos ssize_t len;
507 1.22 christos unsigned char *p, *q, *end;
508 1.22 christos
509 1.22 christos text = a_opt;
510 1.22 christos warn = 0;
511 1.22 christos cr = 0;
512 1.22 christos
513 1.22 christos /* loop over file contents and write to fd */
514 1.22 christos for (int n = 0; ; n++) {
515 1.22 christos if (fd != STDOUT_FILENO)
516 1.22 christos if (tty && (n % 4) == 0)
517 1.22 christos info(" %c\b\b", spinner[(n / 4) % sizeof spinner]);
518 1.22 christos
519 1.22 christos len = archive_read_data(a, buffer, sizeof buffer);
520 1.22 christos
521 1.22 christos if (len < 0)
522 1.22 christos ac(len);
523 1.22 christos
524 1.22 christos /* left over CR from previous buffer */
525 1.22 christos if (a_opt && cr) {
526 1.22 christos if (len == 0 || buffer[0] != '\n')
527 1.22 christos if (write(fd, "\r", 1) != 1)
528 1.22 christos error("write('%s')", pathname);
529 1.22 christos cr = 0;
530 1.22 christos }
531 1.22 christos
532 1.22 christos /* EOF */
533 1.22 christos if (len == 0)
534 1.22 christos break;
535 1.22 christos end = buffer + len;
536 1.22 christos
537 1.22 christos /*
538 1.22 christos * Detect whether this is a text file. The correct way to
539 1.22 christos * do this is to check the least significant bit of the
540 1.22 christos * "internal file attributes" field of the corresponding
541 1.22 christos * file header in the central directory, but libarchive
542 1.22 christos * does not provide access to this field, so we have to
543 1.22 christos * guess by looking for non-ASCII characters in the
544 1.22 christos * buffer. Hopefully we won't guess wrong. If we do
545 1.22 christos * guess wrong, we print a warning message later.
546 1.22 christos */
547 1.22 christos if (a_opt && n == 0) {
548 1.22 christos if (check_binary(buffer, len))
549 1.22 christos text = 0;
550 1.22 christos }
551 1.22 christos
552 1.22 christos /* simple case */
553 1.22 christos if (!a_opt || !text) {
554 1.22 christos if (write(fd, buffer, len) != len)
555 1.22 christos error("write('%s')", pathname);
556 1.22 christos continue;
557 1.22 christos }
558 1.22 christos
559 1.22 christos /* hard case: convert \r\n to \n (sigh...) */
560 1.22 christos for (p = buffer; p < end; p = q + 1) {
561 1.22 christos for (q = p; q < end; q++) {
562 1.22 christos if (!warn && BYTE_IS_BINARY(*q)) {
563 1.22 christos warningx("%s may be corrupted due"
564 1.22 christos " to weak text file detection"
565 1.22 christos " heuristic", pathname);
566 1.22 christos warn = 1;
567 1.22 christos }
568 1.22 christos if (q[0] != '\r')
569 1.22 christos continue;
570 1.22 christos if (&q[1] == end) {
571 1.22 christos cr = 1;
572 1.22 christos break;
573 1.22 christos }
574 1.22 christos if (q[1] == '\n')
575 1.22 christos break;
576 1.22 christos }
577 1.22 christos if (write(fd, p, q - p) != q - p)
578 1.22 christos error("write('%s')", pathname);
579 1.22 christos }
580 1.22 christos }
581 1.22 christos
582 1.22 christos return text;
583 1.22 christos }
584 1.22 christos
585 1.22 christos /*
586 1.1 joerg * Extract a regular file.
587 1.1 joerg */
588 1.1 joerg static void
589 1.8 joerg extract_file(struct archive *a, struct archive_entry *e, char **path)
590 1.1 joerg {
591 1.1 joerg int mode;
592 1.1 joerg time_t mtime;
593 1.1 joerg struct stat sb;
594 1.1 joerg struct timeval tv[2];
595 1.22 christos int fd, check, text;
596 1.20 christos const char *linkname;
597 1.1 joerg
598 1.11 mbalmer mode = archive_entry_mode(e) & 0777;
599 1.1 joerg if (mode == 0)
600 1.1 joerg mode = 0644;
601 1.1 joerg mtime = archive_entry_mtime(e);
602 1.1 joerg
603 1.1 joerg /* look for existing file of same name */
604 1.8 joerg recheck:
605 1.8 joerg if (lstat(*path, &sb) == 0) {
606 1.1 joerg if (u_opt || f_opt) {
607 1.1 joerg /* check if up-to-date */
608 1.1 joerg if (S_ISREG(sb.st_mode) && sb.st_mtime >= mtime)
609 1.1 joerg return;
610 1.8 joerg (void)unlink(*path);
611 1.1 joerg } else if (o_opt) {
612 1.1 joerg /* overwrite */
613 1.8 joerg (void)unlink(*path);
614 1.1 joerg } else if (n_opt) {
615 1.1 joerg /* do not overwrite */
616 1.1 joerg return;
617 1.1 joerg } else {
618 1.8 joerg check = handle_existing_file(path);
619 1.8 joerg if (check == 0)
620 1.8 joerg goto recheck;
621 1.8 joerg if (check == -1)
622 1.8 joerg return; /* do not overwrite */
623 1.1 joerg }
624 1.1 joerg } else {
625 1.1 joerg if (f_opt)
626 1.1 joerg return;
627 1.1 joerg }
628 1.1 joerg
629 1.21 christos tv[0].tv_sec = now;
630 1.21 christos tv[0].tv_usec = 0;
631 1.21 christos tv[1].tv_sec = mtime;
632 1.21 christos tv[1].tv_usec = 0;
633 1.21 christos
634 1.20 christos /* process symlinks */
635 1.20 christos linkname = archive_entry_symlink(e);
636 1.20 christos if (linkname != NULL) {
637 1.20 christos if (symlink(linkname, *path) == -1)
638 1.20 christos error("symlink('%s', '%s')", linkname, *path);
639 1.20 christos info(" extracting: %s -> %s\n", *path, linkname);
640 1.20 christos if (lchmod(*path, mode) == -1)
641 1.20 christos warning("Cannot set mode for '%s'", *path);
642 1.20 christos if (lutimes(*path, tv) == -1)
643 1.20 christos warning("utimes('%s')", *path);
644 1.20 christos return;
645 1.20 christos }
646 1.20 christos
647 1.20 christos /* process hardlinks */
648 1.20 christos linkname = archive_entry_hardlink(e);
649 1.20 christos if (linkname != NULL) {
650 1.20 christos if (link(linkname, *path) == -1)
651 1.20 christos error("link('%s', '%s')", linkname, *path);
652 1.20 christos info(" extracting: %s link to %s\n", *path, linkname);
653 1.20 christos return;
654 1.20 christos }
655 1.20 christos
656 1.8 joerg if ((fd = open(*path, O_RDWR|O_CREAT|O_TRUNC, mode)) < 0)
657 1.8 joerg error("open('%s')", *path);
658 1.1 joerg
659 1.8 joerg info(" extracting: %s", *path);
660 1.1 joerg
661 1.22 christos text = extract2fd(a, *path, fd);
662 1.1 joerg
663 1.1 joerg if (tty)
664 1.1 joerg info(" \b\b");
665 1.1 joerg if (text)
666 1.1 joerg info(" (text)");
667 1.1 joerg info("\n");
668 1.1 joerg
669 1.1 joerg /* set access and modification time */
670 1.1 joerg if (futimes(fd, tv) != 0)
671 1.8 joerg error("utimes('%s')", *path);
672 1.1 joerg if (close(fd) != 0)
673 1.8 joerg error("close('%s')", *path);
674 1.1 joerg }
675 1.1 joerg
676 1.1 joerg /*
677 1.1 joerg * Extract a zipfile entry: first perform some sanity checks to ensure
678 1.1 joerg * that it is either a directory or a regular file and that the path is
679 1.1 joerg * not absolute and does not try to break out of the current directory;
680 1.1 joerg * then call either extract_dir() or extract_file() as appropriate.
681 1.1 joerg *
682 1.1 joerg * This is complicated a bit by the various ways in which we need to
683 1.1 joerg * manipulate the path name. Case conversion (if requested by the -L
684 1.1 joerg * option) happens first, but the include / exclude patterns are applied
685 1.1 joerg * to the full converted path name, before the directory part of the path
686 1.1 joerg * is removed in accordance with the -j option. Sanity checks are
687 1.1 joerg * intentionally done earlier than they need to be, so the user will get a
688 1.1 joerg * warning about insecure paths even for files or directories which
689 1.1 joerg * wouldn't be extracted anyway.
690 1.1 joerg */
691 1.1 joerg static void
692 1.1 joerg extract(struct archive *a, struct archive_entry *e)
693 1.1 joerg {
694 1.1 joerg char *pathname, *realpathname;
695 1.1 joerg mode_t filetype;
696 1.1 joerg char *p, *q;
697 1.1 joerg
698 1.1 joerg pathname = pathdup(archive_entry_pathname(e));
699 1.1 joerg filetype = archive_entry_filetype(e);
700 1.1 joerg
701 1.1 joerg /* sanity checks */
702 1.1 joerg if (pathname[0] == '/' ||
703 1.1 joerg strncmp(pathname, "../", 3) == 0 ||
704 1.1 joerg strstr(pathname, "/../") != NULL) {
705 1.1 joerg warningx("skipping insecure entry '%s'", pathname);
706 1.1 joerg ac(archive_read_data_skip(a));
707 1.1 joerg free(pathname);
708 1.1 joerg return;
709 1.1 joerg }
710 1.1 joerg
711 1.1 joerg /* I don't think this can happen in a zipfile.. */
712 1.20 christos if (!S_ISDIR(filetype) && !S_ISREG(filetype) && !S_ISLNK(filetype)) {
713 1.1 joerg warningx("skipping non-regular entry '%s'", pathname);
714 1.1 joerg ac(archive_read_data_skip(a));
715 1.1 joerg free(pathname);
716 1.1 joerg return;
717 1.1 joerg }
718 1.1 joerg
719 1.1 joerg /* skip directories in -j case */
720 1.1 joerg if (S_ISDIR(filetype) && j_opt) {
721 1.1 joerg ac(archive_read_data_skip(a));
722 1.1 joerg free(pathname);
723 1.1 joerg return;
724 1.1 joerg }
725 1.1 joerg
726 1.1 joerg /* apply include / exclude patterns */
727 1.1 joerg if (!accept_pathname(pathname)) {
728 1.1 joerg ac(archive_read_data_skip(a));
729 1.1 joerg free(pathname);
730 1.1 joerg return;
731 1.1 joerg }
732 1.1 joerg
733 1.1 joerg /* apply -j and -d */
734 1.1 joerg if (j_opt) {
735 1.1 joerg for (p = q = pathname; *p; ++p)
736 1.1 joerg if (*p == '/')
737 1.1 joerg q = p + 1;
738 1.1 joerg realpathname = pathcat(d_arg, q);
739 1.1 joerg } else {
740 1.1 joerg realpathname = pathcat(d_arg, pathname);
741 1.1 joerg }
742 1.1 joerg
743 1.1 joerg /* ensure that parent directory exists */
744 1.1 joerg make_parent(realpathname);
745 1.1 joerg
746 1.1 joerg if (S_ISDIR(filetype))
747 1.1 joerg extract_dir(a, e, realpathname);
748 1.1 joerg else
749 1.8 joerg extract_file(a, e, &realpathname);
750 1.1 joerg
751 1.1 joerg free(realpathname);
752 1.1 joerg free(pathname);
753 1.1 joerg }
754 1.1 joerg
755 1.2 joerg static void
756 1.2 joerg extract_stdout(struct archive *a, struct archive_entry *e)
757 1.2 joerg {
758 1.2 joerg char *pathname;
759 1.2 joerg mode_t filetype;
760 1.2 joerg
761 1.2 joerg pathname = pathdup(archive_entry_pathname(e));
762 1.2 joerg filetype = archive_entry_filetype(e);
763 1.2 joerg
764 1.2 joerg /* I don't think this can happen in a zipfile.. */
765 1.20 christos if (!S_ISDIR(filetype) && !S_ISREG(filetype) && !S_ISLNK(filetype)) {
766 1.2 joerg warningx("skipping non-regular entry '%s'", pathname);
767 1.2 joerg ac(archive_read_data_skip(a));
768 1.2 joerg free(pathname);
769 1.2 joerg return;
770 1.2 joerg }
771 1.2 joerg
772 1.2 joerg /* skip directories in -j case */
773 1.2 joerg if (S_ISDIR(filetype)) {
774 1.2 joerg ac(archive_read_data_skip(a));
775 1.2 joerg free(pathname);
776 1.2 joerg return;
777 1.2 joerg }
778 1.2 joerg
779 1.2 joerg /* apply include / exclude patterns */
780 1.2 joerg if (!accept_pathname(pathname)) {
781 1.2 joerg ac(archive_read_data_skip(a));
782 1.2 joerg free(pathname);
783 1.2 joerg return;
784 1.2 joerg }
785 1.2 joerg
786 1.3 joerg if (c_opt)
787 1.3 joerg info("x %s\n", pathname);
788 1.3 joerg
789 1.22 christos (void)extract2fd(a, pathname, STDOUT_FILENO);
790 1.2 joerg
791 1.2 joerg free(pathname);
792 1.2 joerg }
793 1.2 joerg
794 1.1 joerg /*
795 1.1 joerg * Print the name of an entry to stdout.
796 1.1 joerg */
797 1.1 joerg static void
798 1.1 joerg list(struct archive *a, struct archive_entry *e)
799 1.1 joerg {
800 1.2 joerg char buf[20];
801 1.2 joerg time_t mtime;
802 1.17 christos struct tm *tm;
803 1.1 joerg
804 1.3 joerg mtime = archive_entry_mtime(e);
805 1.17 christos tm = localtime(&mtime);
806 1.17 christos if (*y_str)
807 1.17 christos strftime(buf, sizeof(buf), "%m-%d-%G %R", tm);
808 1.17 christos else
809 1.17 christos strftime(buf, sizeof(buf), "%m-%d-%g %R", tm);
810 1.2 joerg
811 1.3 joerg if (v_opt == 1) {
812 1.3 joerg printf(" %8ju %s %s\n",
813 1.3 joerg (uintmax_t)archive_entry_size(e),
814 1.3 joerg buf, archive_entry_pathname(e));
815 1.3 joerg } else if (v_opt == 2) {
816 1.6 wiz printf("%8ju Stored %7ju 0%% %s %08x %s\n",
817 1.2 joerg (uintmax_t)archive_entry_size(e),
818 1.2 joerg (uintmax_t)archive_entry_size(e),
819 1.2 joerg buf,
820 1.2 joerg 0U,
821 1.2 joerg archive_entry_pathname(e));
822 1.2 joerg }
823 1.1 joerg ac(archive_read_data_skip(a));
824 1.1 joerg }
825 1.1 joerg
826 1.1 joerg /*
827 1.1 joerg * Extract to memory to check CRC
828 1.1 joerg */
829 1.7 wiz static int
830 1.1 joerg test(struct archive *a, struct archive_entry *e)
831 1.1 joerg {
832 1.1 joerg ssize_t len;
833 1.7 wiz int error_count;
834 1.1 joerg
835 1.7 wiz error_count = 0;
836 1.1 joerg if (S_ISDIR(archive_entry_filetype(e)))
837 1.7 wiz return 0;
838 1.1 joerg
839 1.7 wiz info(" testing: %s\t", archive_entry_pathname(e));
840 1.1 joerg while ((len = archive_read_data(a, buffer, sizeof buffer)) > 0)
841 1.1 joerg /* nothing */;
842 1.1 joerg if (len < 0) {
843 1.7 wiz info(" %s\n", archive_error_string(a));
844 1.7 wiz ++error_count;
845 1.1 joerg } else {
846 1.7 wiz info(" OK\n");
847 1.1 joerg }
848 1.1 joerg
849 1.1 joerg /* shouldn't be necessary, but it doesn't hurt */
850 1.1 joerg ac(archive_read_data_skip(a));
851 1.7 wiz
852 1.7 wiz return error_count;
853 1.1 joerg }
854 1.1 joerg
855 1.1 joerg /*
856 1.1 joerg * Main loop: open the zipfile, iterate over its contents and decide what
857 1.1 joerg * to do with each entry.
858 1.1 joerg */
859 1.1 joerg static void
860 1.1 joerg unzip(const char *fn)
861 1.1 joerg {
862 1.1 joerg struct archive *a;
863 1.1 joerg struct archive_entry *e;
864 1.22 christos int ret;
865 1.7 wiz uintmax_t total_size, file_count, error_count;
866 1.1 joerg
867 1.22 christos if ((a = archive_read_new()) == NULL)
868 1.22 christos error("archive_read_new failed");
869 1.1 joerg
870 1.1 joerg ac(archive_read_support_format_zip(a));
871 1.22 christos ac(archive_read_open_filename(a, fn, 8192));
872 1.1 joerg
873 1.12 joerg if (!q_opt && !p_opt)
874 1.9 wiz printf("Archive: %s\n", fn);
875 1.9 wiz
876 1.3 joerg if (v_opt == 1) {
877 1.17 christos printf(" Length %sDate Time Name\n", y_str);
878 1.17 christos printf(" -------- %s---- ---- ----\n", y_str);
879 1.3 joerg } else if (v_opt == 2) {
880 1.17 christos printf(" Length Method Size Ratio %sDate Time CRC-32 Name\n", y_str);
881 1.17 christos printf("-------- ------ ------- ----- %s---- ---- ------ ----\n", y_str);
882 1.3 joerg }
883 1.3 joerg
884 1.3 joerg total_size = 0;
885 1.3 joerg file_count = 0;
886 1.7 wiz error_count = 0;
887 1.1 joerg for (;;) {
888 1.1 joerg ret = archive_read_next_header(a, &e);
889 1.1 joerg if (ret == ARCHIVE_EOF)
890 1.1 joerg break;
891 1.1 joerg ac(ret);
892 1.1 joerg if (t_opt)
893 1.7 wiz error_count += test(a, e);
894 1.2 joerg else if (v_opt)
895 1.1 joerg list(a, e);
896 1.3 joerg else if (p_opt || c_opt)
897 1.2 joerg extract_stdout(a, e);
898 1.1 joerg else
899 1.1 joerg extract(a, e);
900 1.3 joerg
901 1.3 joerg total_size += archive_entry_size(e);
902 1.3 joerg ++file_count;
903 1.3 joerg }
904 1.3 joerg
905 1.3 joerg if (v_opt == 1) {
906 1.17 christos printf(" -------- %s-------\n", y_str);
907 1.17 christos printf(" %8ju %s%ju file%s\n",
908 1.17 christos total_size, y_str, file_count, file_count != 1 ? "s" : "");
909 1.3 joerg } else if (v_opt == 2) {
910 1.17 christos printf("-------- ------- --- %s-------\n", y_str);
911 1.17 christos printf("%8ju %7ju 0%% %s%ju file%s\n",
912 1.17 christos total_size, total_size, y_str, file_count,
913 1.3 joerg file_count != 1 ? "s" : "");
914 1.1 joerg }
915 1.1 joerg
916 1.23 joerg ac(archive_read_free(a));
917 1.7 wiz
918 1.7 wiz if (t_opt) {
919 1.7 wiz if (error_count > 0) {
920 1.19 joerg errorx("%ju checksum error(s) found.", error_count);
921 1.7 wiz }
922 1.7 wiz else {
923 1.7 wiz printf("No errors detected in compressed data of %s.\n",
924 1.7 wiz fn);
925 1.7 wiz }
926 1.7 wiz }
927 1.1 joerg }
928 1.1 joerg
929 1.17 christos static void __dead
930 1.1 joerg usage(void)
931 1.1 joerg {
932 1.1 joerg
933 1.17 christos fprintf(stderr, "Usage: %s [-aCcfjLlnopqtuvy] [-d dir] [-x pattern] "
934 1.17 christos "zipfile\n", getprogname());
935 1.1 joerg exit(1);
936 1.1 joerg }
937 1.1 joerg
938 1.1 joerg static int
939 1.1 joerg getopts(int argc, char *argv[])
940 1.1 joerg {
941 1.1 joerg int opt;
942 1.1 joerg
943 1.25 christos #ifdef __GLIBC__
944 1.25 christos optind = 0;
945 1.25 christos #else
946 1.25 christos optreset = optind = 1;
947 1.25 christos #endif
948 1.25 christos
949 1.25 christos while ((opt = getopt(argc, argv, "aCcd:fjLlnopP:qtuvyx:")) != -1)
950 1.17 christos while ((opt = getopt(argc, argv, "aCcd:fjLlnopqtuvyx:")) != -1)
951 1.1 joerg switch (opt) {
952 1.1 joerg case 'a':
953 1.1 joerg a_opt = 1;
954 1.1 joerg break;
955 1.7 wiz case 'C':
956 1.7 wiz C_opt = 1;
957 1.7 wiz break;
958 1.3 joerg case 'c':
959 1.3 joerg c_opt = 1;
960 1.3 joerg break;
961 1.1 joerg case 'd':
962 1.1 joerg d_arg = optarg;
963 1.1 joerg break;
964 1.1 joerg case 'f':
965 1.1 joerg f_opt = 1;
966 1.1 joerg break;
967 1.1 joerg case 'j':
968 1.1 joerg j_opt = 1;
969 1.1 joerg break;
970 1.1 joerg case 'L':
971 1.1 joerg L_opt = 1;
972 1.1 joerg break;
973 1.1 joerg case 'l':
974 1.2 joerg if (v_opt == 0)
975 1.2 joerg v_opt = 1;
976 1.1 joerg break;
977 1.1 joerg case 'n':
978 1.1 joerg n_opt = 1;
979 1.1 joerg break;
980 1.1 joerg case 'o':
981 1.1 joerg o_opt = 1;
982 1.2 joerg q_opt = 1;
983 1.2 joerg break;
984 1.2 joerg case 'p':
985 1.2 joerg p_opt = 1;
986 1.1 joerg break;
987 1.1 joerg case 'q':
988 1.1 joerg q_opt = 1;
989 1.1 joerg break;
990 1.1 joerg case 't':
991 1.1 joerg t_opt = 1;
992 1.1 joerg break;
993 1.1 joerg case 'u':
994 1.1 joerg u_opt = 1;
995 1.1 joerg break;
996 1.2 joerg case 'v':
997 1.2 joerg v_opt = 2;
998 1.2 joerg break;
999 1.1 joerg case 'x':
1000 1.1 joerg add_pattern(&exclude, optarg);
1001 1.1 joerg break;
1002 1.17 christos case 'y':
1003 1.17 christos y_str = " ";
1004 1.17 christos break;
1005 1.1 joerg default:
1006 1.1 joerg usage();
1007 1.1 joerg }
1008 1.1 joerg
1009 1.1 joerg return (optind);
1010 1.1 joerg }
1011 1.1 joerg
1012 1.1 joerg int
1013 1.1 joerg main(int argc, char *argv[])
1014 1.1 joerg {
1015 1.1 joerg const char *zipfile;
1016 1.1 joerg int nopts;
1017 1.1 joerg
1018 1.1 joerg if (isatty(STDOUT_FILENO))
1019 1.1 joerg tty = 1;
1020 1.1 joerg
1021 1.1 joerg if (getenv("UNZIP_DEBUG") != NULL)
1022 1.1 joerg unzip_debug = 1;
1023 1.1 joerg for (int i = 0; i < argc; ++i)
1024 1.1 joerg debug("%s%c", argv[i], (i < argc - 1) ? ' ' : '\n');
1025 1.1 joerg
1026 1.1 joerg /*
1027 1.1 joerg * Info-ZIP's unzip(1) expects certain options to come before the
1028 1.1 joerg * zipfile name, and others to come after - though it does not
1029 1.1 joerg * enforce this. For simplicity, we accept *all* options both
1030 1.1 joerg * before and after the zipfile name.
1031 1.1 joerg */
1032 1.1 joerg nopts = getopts(argc, argv);
1033 1.1 joerg
1034 1.1 joerg if (argc <= nopts)
1035 1.1 joerg usage();
1036 1.1 joerg zipfile = argv[nopts++];
1037 1.1 joerg
1038 1.22 christos if (strcmp(zipfile, "-") == 0)
1039 1.22 christos zipfile = NULL; /* STDIN */
1040 1.22 christos
1041 1.1 joerg while (nopts < argc && *argv[nopts] != '-')
1042 1.1 joerg add_pattern(&include, argv[nopts++]);
1043 1.1 joerg
1044 1.1 joerg nopts--; /* fake argv[0] */
1045 1.1 joerg nopts += getopts(argc - nopts, argv + nopts);
1046 1.1 joerg
1047 1.1 joerg if (n_opt + o_opt + u_opt > 1)
1048 1.1 joerg errorx("-n, -o and -u are contradictory");
1049 1.1 joerg
1050 1.1 joerg time(&now);
1051 1.1 joerg
1052 1.1 joerg unzip(zipfile);
1053 1.1 joerg
1054 1.1 joerg exit(0);
1055 1.1 joerg }
1056