unzip.c revision 1.23.10.1 1 /* $NetBSD: unzip.c,v 1.23.10.1 2018/07/28 04:38:14 pgoyette Exp $ */
2
3 /*-
4 * Copyright (c) 2009, 2010 Joerg Sonnenberger <joerg (at) NetBSD.org>
5 * Copyright (c) 2007-2008 Dag-Erling Codan Smrgrav
6 * All rights reserved.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer
13 * in this position and unchanged.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 *
18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
22 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 * SUCH DAMAGE.
29 *
30 * $FreeBSD: revision 180124$
31 *
32 * This file would be much shorter if we didn't care about command-line
33 * compatibility with Info-ZIP's UnZip, which requires us to duplicate
34 * parts of libarchive in order to gain more detailed control of its
35 * behaviour for the purpose of implementing the -n, -o, -L and -a
36 * options.
37 */
38
39 #include <sys/cdefs.h>
40 __RCSID("$NetBSD: unzip.c,v 1.23.10.1 2018/07/28 04:38:14 pgoyette Exp $");
41
42 #include <sys/queue.h>
43 #include <sys/stat.h>
44
45 #include <ctype.h>
46 #include <errno.h>
47 #include <fcntl.h>
48 #include <fnmatch.h>
49 #include <stdarg.h>
50 #include <stdio.h>
51 #include <stdlib.h>
52 #include <string.h>
53 #include <unistd.h>
54
55 #include <archive.h>
56 #include <archive_entry.h>
57
58 /* command-line options */
59 static int a_opt; /* convert EOL */
60 static int C_opt; /* match case-insensitively */
61 static int c_opt; /* extract to stdout */
62 static const char *d_arg; /* directory */
63 static int f_opt; /* update existing files only */
64 static int j_opt; /* junk directories */
65 static int L_opt; /* lowercase names */
66 static int n_opt; /* never overwrite */
67 static int o_opt; /* always overwrite */
68 static int p_opt; /* extract to stdout, quiet */
69 static int q_opt; /* quiet */
70 static int t_opt; /* test */
71 static int u_opt; /* update */
72 static int v_opt; /* verbose/list */
73 static const char * y_str = ""; /* 4 digit year */
74
75 /* time when unzip started */
76 static time_t now;
77
78 /* debug flag */
79 static int unzip_debug;
80
81 /* running on tty? */
82 static int tty;
83
84 /* convenience macro */
85 /* XXX should differentiate between ARCHIVE_{WARN,FAIL,RETRY} */
86 #define ac(call) \
87 do { \
88 int acret = (call); \
89 if (acret != ARCHIVE_OK) \
90 errorx("%s", archive_error_string(a)); \
91 } while (0)
92
93 /*
94 * Indicates that last info() did not end with EOL. This helps error() et
95 * al. avoid printing an error message on the same line as an incomplete
96 * informational message.
97 */
98 static int noeol;
99
100 /* fatal error message + errno */
101 __dead __printflike(1, 2) static void
102 error(const char *fmt, ...)
103 {
104 va_list ap;
105
106 if (noeol)
107 fprintf(stdout, "\n");
108 fflush(stdout);
109 fprintf(stderr, "unzip: ");
110 va_start(ap, fmt);
111 vfprintf(stderr, fmt, ap);
112 va_end(ap);
113 fprintf(stderr, ": %s\n", strerror(errno));
114 exit(1);
115 }
116
117 /* fatal error message, no errno */
118 __dead __printflike(1, 2) static void
119 errorx(const char *fmt, ...)
120 {
121 va_list ap;
122
123 if (noeol)
124 fprintf(stdout, "\n");
125 fflush(stdout);
126 fprintf(stderr, "unzip: ");
127 va_start(ap, fmt);
128 vfprintf(stderr, fmt, ap);
129 va_end(ap);
130 fprintf(stderr, "\n");
131 exit(1);
132 }
133
134 /* non-fatal error message + errno */
135 __printflike(1, 2) static void
136 warning(const char *fmt, ...)
137 {
138 va_list ap;
139
140 if (noeol)
141 fprintf(stdout, "\n");
142 fflush(stdout);
143 fprintf(stderr, "unzip: ");
144 va_start(ap, fmt);
145 vfprintf(stderr, fmt, ap);
146 va_end(ap);
147 fprintf(stderr, ": %s\n", strerror(errno));
148 }
149
150 /* non-fatal error message, no errno */
151 __printflike(1, 2) static void
152 warningx(const char *fmt, ...)
153 {
154 va_list ap;
155
156 if (noeol)
157 fprintf(stdout, "\n");
158 fflush(stdout);
159 fprintf(stderr, "unzip: ");
160 va_start(ap, fmt);
161 vfprintf(stderr, fmt, ap);
162 va_end(ap);
163 fprintf(stderr, "\n");
164 }
165
166 /* informational message (if not -q) */
167 __printflike(1, 2) static void
168 info(const char *fmt, ...)
169 {
170 va_list ap;
171
172 if (q_opt && !unzip_debug)
173 return;
174 va_start(ap, fmt);
175 vfprintf(stdout, fmt, ap);
176 va_end(ap);
177 fflush(stdout);
178
179 if (*fmt == '\0')
180 noeol = 1;
181 else
182 noeol = fmt[strlen(fmt) - 1] != '\n';
183 }
184
185 /* debug message (if unzip_debug) */
186 __printflike(1, 2) static void
187 debug(const char *fmt, ...)
188 {
189 va_list ap;
190
191 if (!unzip_debug)
192 return;
193 va_start(ap, fmt);
194 vfprintf(stderr, fmt, ap);
195 va_end(ap);
196 fflush(stderr);
197
198 if (*fmt == '\0')
199 noeol = 1;
200 else
201 noeol = fmt[strlen(fmt) - 1] != '\n';
202 }
203
204 /* duplicate a path name, possibly converting to lower case */
205 static char *
206 pathdup(const char *path)
207 {
208 char *str;
209 size_t i, len;
210
211 len = strlen(path);
212 while (len && path[len - 1] == '/')
213 len--;
214 if ((str = malloc(len + 1)) == NULL) {
215 errno = ENOMEM;
216 error("malloc()");
217 }
218 if (L_opt) {
219 for (i = 0; i < len; ++i)
220 str[i] = tolower((unsigned char)path[i]);
221 } else {
222 memcpy(str, path, len);
223 }
224 str[len] = '\0';
225
226 return (str);
227 }
228
229 /* concatenate two path names */
230 static char *
231 pathcat(const char *prefix, const char *path)
232 {
233 char *str;
234 size_t prelen, len;
235
236 prelen = prefix ? strlen(prefix) + 1 : 0;
237 len = strlen(path) + 1;
238 if ((str = malloc(prelen + len)) == NULL) {
239 errno = ENOMEM;
240 error("malloc()");
241 }
242 if (prefix) {
243 memcpy(str, prefix, prelen); /* includes zero */
244 str[prelen - 1] = '/'; /* splat zero */
245 }
246 memcpy(str + prelen, path, len); /* includes zero */
247
248 return (str);
249 }
250
251 /*
252 * Pattern lists for include / exclude processing
253 */
254 struct pattern {
255 STAILQ_ENTRY(pattern) link;
256 char pattern[];
257 };
258
259 STAILQ_HEAD(pattern_list, pattern);
260 static struct pattern_list include = STAILQ_HEAD_INITIALIZER(include);
261 static struct pattern_list exclude = STAILQ_HEAD_INITIALIZER(exclude);
262
263 /*
264 * Add an entry to a pattern list
265 */
266 static void
267 add_pattern(struct pattern_list *list, const char *pattern)
268 {
269 struct pattern *entry;
270 size_t len;
271
272 debug("adding pattern '%s'\n", pattern);
273 len = strlen(pattern);
274 if ((entry = malloc(sizeof *entry + len + 1)) == NULL) {
275 errno = ENOMEM;
276 error("malloc()");
277 }
278 memcpy(entry->pattern, pattern, len + 1);
279 STAILQ_INSERT_TAIL(list, entry, link);
280 }
281
282 /*
283 * Match a string against a list of patterns
284 */
285 static int
286 match_pattern(struct pattern_list *list, const char *str)
287 {
288 struct pattern *entry;
289
290 STAILQ_FOREACH(entry, list, link) {
291 if (fnmatch(entry->pattern, str, C_opt ? FNM_CASEFOLD : 0) == 0)
292 return (1);
293 }
294 return (0);
295 }
296
297 /*
298 * Verify that a given pathname is in the include list and not in the
299 * exclude list.
300 */
301 static int
302 accept_pathname(const char *pathname)
303 {
304
305 if (!STAILQ_EMPTY(&include) && !match_pattern(&include, pathname))
306 return (0);
307 if (!STAILQ_EMPTY(&exclude) && match_pattern(&exclude, pathname))
308 return (0);
309 return (1);
310 }
311
312 /*
313 * Create the specified directory with the specified mode, taking certain
314 * precautions on they way.
315 */
316 static void
317 make_dir(const char *path, int mode)
318 {
319 struct stat sb;
320
321 if (lstat(path, &sb) == 0) {
322 if (S_ISDIR(sb.st_mode))
323 return;
324 /*
325 * Normally, we should either ask the user about removing
326 * the non-directory of the same name as a directory we
327 * wish to create, or respect the -n or -o command-line
328 * options. However, this may lead to a later failure or
329 * even compromise (if this non-directory happens to be a
330 * symlink to somewhere unsafe), so we don't.
331 */
332
333 /*
334 * Don't check unlink() result; failure will cause mkdir()
335 * to fail later, which we will catch.
336 */
337 (void)unlink(path);
338 }
339 if (mkdir(path, mode) != 0 && errno != EEXIST)
340 error("mkdir('%s')", path);
341 }
342
343 /*
344 * Ensure that all directories leading up to (but not including) the
345 * specified path exist.
346 *
347 * XXX inefficient + modifies the file in-place
348 */
349 static void
350 make_parent(char *path)
351 {
352 struct stat sb;
353 char *sep;
354
355 sep = strrchr(path, '/');
356 if (sep == NULL || sep == path)
357 return;
358 *sep = '\0';
359 if (lstat(path, &sb) == 0) {
360 if (S_ISDIR(sb.st_mode)) {
361 *sep = '/';
362 return;
363 }
364 unlink(path);
365 }
366 make_parent(path);
367 mkdir(path, 0755);
368 *sep = '/';
369
370 #if 0
371 for (sep = path; (sep = strchr(sep, '/')) != NULL; sep++) {
372 /* root in case of absolute d_arg */
373 if (sep == path)
374 continue;
375 *sep = '\0';
376 make_dir(path, 0755);
377 *sep = '/';
378 }
379 #endif
380 }
381
382 /*
383 * Extract a directory.
384 */
385 static void
386 extract_dir(struct archive *a, struct archive_entry *e, const char *path)
387 {
388 int mode;
389
390 /*
391 * Dropbox likes to create '/' directory entries, just ignore
392 * such junk.
393 */
394 if (*path == '\0')
395 return;
396
397 mode = archive_entry_mode(e) & 0777;
398 if (mode == 0)
399 mode = 0755;
400
401 /*
402 * Some zipfiles contain directories with weird permissions such
403 * as 0644 or 0444. This can cause strange issues such as being
404 * unable to extract files into the directory we just created, or
405 * the user being unable to remove the directory later without
406 * first manually changing its permissions. Therefore, we whack
407 * the permissions into shape, assuming that the user wants full
408 * access and that anyone who gets read access also gets execute
409 * access.
410 */
411 mode |= 0700;
412 if (mode & 0040)
413 mode |= 0010;
414 if (mode & 0004)
415 mode |= 0001;
416
417 info(" creating: %s/\n", path);
418 make_dir(path, mode);
419 ac(archive_read_data_skip(a));
420 }
421
422 static unsigned char buffer[8192];
423 static char spinner[] = { '|', '/', '-', '\\' };
424
425 static int
426 handle_existing_file(char **path)
427 {
428 size_t alen;
429 ssize_t len;
430 char buf[4];
431
432 for (;;) {
433 fprintf(stderr,
434 "replace %s? [y]es, [n]o, [A]ll, [N]one, [r]ename: ",
435 *path);
436 fgets(buf, 4, stdin);
437 switch (*buf) {
438 case 'A':
439 o_opt = 1;
440 /* FALL THROUGH */
441 case 'y':
442 case 'Y':
443 (void)unlink(*path);
444 return 1;
445 case 'N':
446 n_opt = 1;
447 /* FALL THROUGH */
448 case 'n':
449 return -1;
450 case 'r':
451 case 'R':
452 printf("New name: ");
453 fflush(stdout);
454 free(*path);
455 *path = NULL;
456 alen = 0;
457 len = getline(path, &alen, stdin);
458 if ((*path)[len - 1] == '\n')
459 (*path)[len - 1] = '\0';
460 return 0;
461 default:
462 break;
463 }
464 }
465 }
466
467 /*
468 * Detect binary files by a combination of character white list and
469 * black list. NUL bytes and other control codes without use in text files
470 * result directly in switching the file to binary mode. Otherwise, at least
471 * one white-listed byte has to be found.
472 *
473 * Black-listed: 0..6, 14..25, 28..31
474 * White-listed: 9..10, 13, >= 32
475 *
476 * See the proginfo/txtvsbin.txt in the zip sources for a detailed discussion.
477 */
478 #define BYTE_IS_BINARY(x) ((x) < 32 && (0xf3ffc07fU & (1U << (x))))
479 #define BYTE_IS_TEXT(x) ((x) >= 32 || (0x00002600U & (1U << (x))))
480
481 static int
482 check_binary(const unsigned char *buf, size_t len)
483 {
484 int rv;
485 for (rv = 1; len--; ++buf) {
486 if (BYTE_IS_BINARY(*buf))
487 return 1;
488 if (BYTE_IS_TEXT(*buf))
489 rv = 0;
490 }
491
492 return rv;
493 }
494
495 /*
496 * Extract to a file descriptor
497 */
498 static int
499 extract2fd(struct archive *a, char *pathname, int fd)
500 {
501 int cr, text, warn;
502 ssize_t len;
503 unsigned char *p, *q, *end;
504
505 text = a_opt;
506 warn = 0;
507 cr = 0;
508
509 /* loop over file contents and write to fd */
510 for (int n = 0; ; n++) {
511 if (fd != STDOUT_FILENO)
512 if (tty && (n % 4) == 0)
513 info(" %c\b\b", spinner[(n / 4) % sizeof spinner]);
514
515 len = archive_read_data(a, buffer, sizeof buffer);
516
517 if (len < 0)
518 ac(len);
519
520 /* left over CR from previous buffer */
521 if (a_opt && cr) {
522 if (len == 0 || buffer[0] != '\n')
523 if (write(fd, "\r", 1) != 1)
524 error("write('%s')", pathname);
525 cr = 0;
526 }
527
528 /* EOF */
529 if (len == 0)
530 break;
531 end = buffer + len;
532
533 /*
534 * Detect whether this is a text file. The correct way to
535 * do this is to check the least significant bit of the
536 * "internal file attributes" field of the corresponding
537 * file header in the central directory, but libarchive
538 * does not provide access to this field, so we have to
539 * guess by looking for non-ASCII characters in the
540 * buffer. Hopefully we won't guess wrong. If we do
541 * guess wrong, we print a warning message later.
542 */
543 if (a_opt && n == 0) {
544 if (check_binary(buffer, len))
545 text = 0;
546 }
547
548 /* simple case */
549 if (!a_opt || !text) {
550 if (write(fd, buffer, len) != len)
551 error("write('%s')", pathname);
552 continue;
553 }
554
555 /* hard case: convert \r\n to \n (sigh...) */
556 for (p = buffer; p < end; p = q + 1) {
557 for (q = p; q < end; q++) {
558 if (!warn && BYTE_IS_BINARY(*q)) {
559 warningx("%s may be corrupted due"
560 " to weak text file detection"
561 " heuristic", pathname);
562 warn = 1;
563 }
564 if (q[0] != '\r')
565 continue;
566 if (&q[1] == end) {
567 cr = 1;
568 break;
569 }
570 if (q[1] == '\n')
571 break;
572 }
573 if (write(fd, p, q - p) != q - p)
574 error("write('%s')", pathname);
575 }
576 }
577
578 return text;
579 }
580
581 /*
582 * Extract a regular file.
583 */
584 static void
585 extract_file(struct archive *a, struct archive_entry *e, char **path)
586 {
587 int mode;
588 time_t mtime;
589 struct stat sb;
590 struct timeval tv[2];
591 int fd, check, text;
592 const char *linkname;
593
594 mode = archive_entry_mode(e) & 0777;
595 if (mode == 0)
596 mode = 0644;
597 mtime = archive_entry_mtime(e);
598
599 /* look for existing file of same name */
600 recheck:
601 if (lstat(*path, &sb) == 0) {
602 if (u_opt || f_opt) {
603 /* check if up-to-date */
604 if (S_ISREG(sb.st_mode) && sb.st_mtime >= mtime)
605 return;
606 (void)unlink(*path);
607 } else if (o_opt) {
608 /* overwrite */
609 (void)unlink(*path);
610 } else if (n_opt) {
611 /* do not overwrite */
612 return;
613 } else {
614 check = handle_existing_file(path);
615 if (check == 0)
616 goto recheck;
617 if (check == -1)
618 return; /* do not overwrite */
619 }
620 } else {
621 if (f_opt)
622 return;
623 }
624
625 tv[0].tv_sec = now;
626 tv[0].tv_usec = 0;
627 tv[1].tv_sec = mtime;
628 tv[1].tv_usec = 0;
629
630 /* process symlinks */
631 linkname = archive_entry_symlink(e);
632 if (linkname != NULL) {
633 if (symlink(linkname, *path) == -1)
634 error("symlink('%s', '%s')", linkname, *path);
635 info(" extracting: %s -> %s\n", *path, linkname);
636 if (lchmod(*path, mode) == -1)
637 warning("Cannot set mode for '%s'", *path);
638 if (lutimes(*path, tv) == -1)
639 warning("utimes('%s')", *path);
640 return;
641 }
642
643 /* process hardlinks */
644 linkname = archive_entry_hardlink(e);
645 if (linkname != NULL) {
646 if (link(linkname, *path) == -1)
647 error("link('%s', '%s')", linkname, *path);
648 info(" extracting: %s link to %s\n", *path, linkname);
649 return;
650 }
651
652 if ((fd = open(*path, O_RDWR|O_CREAT|O_TRUNC, mode)) < 0)
653 error("open('%s')", *path);
654
655 info(" extracting: %s", *path);
656
657 text = extract2fd(a, *path, fd);
658
659 if (tty)
660 info(" \b\b");
661 if (text)
662 info(" (text)");
663 info("\n");
664
665 /* set access and modification time */
666 if (futimes(fd, tv) != 0)
667 error("utimes('%s')", *path);
668 if (close(fd) != 0)
669 error("close('%s')", *path);
670 }
671
672 /*
673 * Extract a zipfile entry: first perform some sanity checks to ensure
674 * that it is either a directory or a regular file and that the path is
675 * not absolute and does not try to break out of the current directory;
676 * then call either extract_dir() or extract_file() as appropriate.
677 *
678 * This is complicated a bit by the various ways in which we need to
679 * manipulate the path name. Case conversion (if requested by the -L
680 * option) happens first, but the include / exclude patterns are applied
681 * to the full converted path name, before the directory part of the path
682 * is removed in accordance with the -j option. Sanity checks are
683 * intentionally done earlier than they need to be, so the user will get a
684 * warning about insecure paths even for files or directories which
685 * wouldn't be extracted anyway.
686 */
687 static void
688 extract(struct archive *a, struct archive_entry *e)
689 {
690 char *pathname, *realpathname;
691 mode_t filetype;
692 char *p, *q;
693
694 pathname = pathdup(archive_entry_pathname(e));
695 filetype = archive_entry_filetype(e);
696
697 /* sanity checks */
698 if (pathname[0] == '/' ||
699 strncmp(pathname, "../", 3) == 0 ||
700 strstr(pathname, "/../") != NULL) {
701 warningx("skipping insecure entry '%s'", pathname);
702 ac(archive_read_data_skip(a));
703 free(pathname);
704 return;
705 }
706
707 /* I don't think this can happen in a zipfile.. */
708 if (!S_ISDIR(filetype) && !S_ISREG(filetype) && !S_ISLNK(filetype)) {
709 warningx("skipping non-regular entry '%s'", pathname);
710 ac(archive_read_data_skip(a));
711 free(pathname);
712 return;
713 }
714
715 /* skip directories in -j case */
716 if (S_ISDIR(filetype) && j_opt) {
717 ac(archive_read_data_skip(a));
718 free(pathname);
719 return;
720 }
721
722 /* apply include / exclude patterns */
723 if (!accept_pathname(pathname)) {
724 ac(archive_read_data_skip(a));
725 free(pathname);
726 return;
727 }
728
729 /* apply -j and -d */
730 if (j_opt) {
731 for (p = q = pathname; *p; ++p)
732 if (*p == '/')
733 q = p + 1;
734 realpathname = pathcat(d_arg, q);
735 } else {
736 realpathname = pathcat(d_arg, pathname);
737 }
738
739 /* ensure that parent directory exists */
740 make_parent(realpathname);
741
742 if (S_ISDIR(filetype))
743 extract_dir(a, e, realpathname);
744 else
745 extract_file(a, e, &realpathname);
746
747 free(realpathname);
748 free(pathname);
749 }
750
751 static void
752 extract_stdout(struct archive *a, struct archive_entry *e)
753 {
754 char *pathname;
755 mode_t filetype;
756
757 pathname = pathdup(archive_entry_pathname(e));
758 filetype = archive_entry_filetype(e);
759
760 /* I don't think this can happen in a zipfile.. */
761 if (!S_ISDIR(filetype) && !S_ISREG(filetype) && !S_ISLNK(filetype)) {
762 warningx("skipping non-regular entry '%s'", pathname);
763 ac(archive_read_data_skip(a));
764 free(pathname);
765 return;
766 }
767
768 /* skip directories in -j case */
769 if (S_ISDIR(filetype)) {
770 ac(archive_read_data_skip(a));
771 free(pathname);
772 return;
773 }
774
775 /* apply include / exclude patterns */
776 if (!accept_pathname(pathname)) {
777 ac(archive_read_data_skip(a));
778 free(pathname);
779 return;
780 }
781
782 if (c_opt)
783 info("x %s\n", pathname);
784
785 (void)extract2fd(a, pathname, STDOUT_FILENO);
786
787 free(pathname);
788 }
789
790 /*
791 * Print the name of an entry to stdout.
792 */
793 static void
794 list(struct archive *a, struct archive_entry *e)
795 {
796 char buf[20];
797 time_t mtime;
798 struct tm *tm;
799
800 mtime = archive_entry_mtime(e);
801 tm = localtime(&mtime);
802 if (*y_str)
803 strftime(buf, sizeof(buf), "%m-%d-%G %R", tm);
804 else
805 strftime(buf, sizeof(buf), "%m-%d-%g %R", tm);
806
807 if (v_opt == 1) {
808 printf(" %8ju %s %s\n",
809 (uintmax_t)archive_entry_size(e),
810 buf, archive_entry_pathname(e));
811 } else if (v_opt == 2) {
812 printf("%8ju Stored %7ju 0%% %s %08x %s\n",
813 (uintmax_t)archive_entry_size(e),
814 (uintmax_t)archive_entry_size(e),
815 buf,
816 0U,
817 archive_entry_pathname(e));
818 }
819 ac(archive_read_data_skip(a));
820 }
821
822 /*
823 * Extract to memory to check CRC
824 */
825 static int
826 test(struct archive *a, struct archive_entry *e)
827 {
828 ssize_t len;
829 int error_count;
830
831 error_count = 0;
832 if (S_ISDIR(archive_entry_filetype(e)))
833 return 0;
834
835 info(" testing: %s\t", archive_entry_pathname(e));
836 while ((len = archive_read_data(a, buffer, sizeof buffer)) > 0)
837 /* nothing */;
838 if (len < 0) {
839 info(" %s\n", archive_error_string(a));
840 ++error_count;
841 } else {
842 info(" OK\n");
843 }
844
845 /* shouldn't be necessary, but it doesn't hurt */
846 ac(archive_read_data_skip(a));
847
848 return error_count;
849 }
850
851 /*
852 * Main loop: open the zipfile, iterate over its contents and decide what
853 * to do with each entry.
854 */
855 static void
856 unzip(const char *fn)
857 {
858 struct archive *a;
859 struct archive_entry *e;
860 int ret;
861 uintmax_t total_size, file_count, error_count;
862
863 if ((a = archive_read_new()) == NULL)
864 error("archive_read_new failed");
865
866 ac(archive_read_support_format_zip(a));
867 ac(archive_read_open_filename(a, fn, 8192));
868
869 if (!q_opt && !p_opt)
870 printf("Archive: %s\n", fn);
871
872 if (v_opt == 1) {
873 printf(" Length %sDate Time Name\n", y_str);
874 printf(" -------- %s---- ---- ----\n", y_str);
875 } else if (v_opt == 2) {
876 printf(" Length Method Size Ratio %sDate Time CRC-32 Name\n", y_str);
877 printf("-------- ------ ------- ----- %s---- ---- ------ ----\n", y_str);
878 }
879
880 total_size = 0;
881 file_count = 0;
882 error_count = 0;
883 for (;;) {
884 ret = archive_read_next_header(a, &e);
885 if (ret == ARCHIVE_EOF)
886 break;
887 ac(ret);
888 if (t_opt)
889 error_count += test(a, e);
890 else if (v_opt)
891 list(a, e);
892 else if (p_opt || c_opt)
893 extract_stdout(a, e);
894 else
895 extract(a, e);
896
897 total_size += archive_entry_size(e);
898 ++file_count;
899 }
900
901 if (v_opt == 1) {
902 printf(" -------- %s-------\n", y_str);
903 printf(" %8ju %s%ju file%s\n",
904 total_size, y_str, file_count, file_count != 1 ? "s" : "");
905 } else if (v_opt == 2) {
906 printf("-------- ------- --- %s-------\n", y_str);
907 printf("%8ju %7ju 0%% %s%ju file%s\n",
908 total_size, total_size, y_str, file_count,
909 file_count != 1 ? "s" : "");
910 }
911
912 ac(archive_read_free(a));
913
914 if (t_opt) {
915 if (error_count > 0) {
916 errorx("%ju checksum error(s) found.", error_count);
917 }
918 else {
919 printf("No errors detected in compressed data of %s.\n",
920 fn);
921 }
922 }
923 }
924
925 static void __dead
926 usage(void)
927 {
928
929 fprintf(stderr, "Usage: %s [-aCcfjLlnopqtuvy] [-d dir] [-x pattern] "
930 "zipfile\n", getprogname());
931 exit(1);
932 }
933
934 static int
935 getopts(int argc, char *argv[])
936 {
937 int opt;
938
939 optreset = optind = 1;
940 while ((opt = getopt(argc, argv, "aCcd:fjLlnopqtuvyx:")) != -1)
941 switch (opt) {
942 case 'a':
943 a_opt = 1;
944 break;
945 case 'C':
946 C_opt = 1;
947 break;
948 case 'c':
949 c_opt = 1;
950 break;
951 case 'd':
952 d_arg = optarg;
953 break;
954 case 'f':
955 f_opt = 1;
956 break;
957 case 'j':
958 j_opt = 1;
959 break;
960 case 'L':
961 L_opt = 1;
962 break;
963 case 'l':
964 if (v_opt == 0)
965 v_opt = 1;
966 break;
967 case 'n':
968 n_opt = 1;
969 break;
970 case 'o':
971 o_opt = 1;
972 q_opt = 1;
973 break;
974 case 'p':
975 p_opt = 1;
976 break;
977 case 'q':
978 q_opt = 1;
979 break;
980 case 't':
981 t_opt = 1;
982 break;
983 case 'u':
984 u_opt = 1;
985 break;
986 case 'v':
987 v_opt = 2;
988 break;
989 case 'x':
990 add_pattern(&exclude, optarg);
991 break;
992 case 'y':
993 y_str = " ";
994 break;
995 default:
996 usage();
997 }
998
999 return (optind);
1000 }
1001
1002 int
1003 main(int argc, char *argv[])
1004 {
1005 const char *zipfile;
1006 int nopts;
1007
1008 if (isatty(STDOUT_FILENO))
1009 tty = 1;
1010
1011 if (getenv("UNZIP_DEBUG") != NULL)
1012 unzip_debug = 1;
1013 for (int i = 0; i < argc; ++i)
1014 debug("%s%c", argv[i], (i < argc - 1) ? ' ' : '\n');
1015
1016 /*
1017 * Info-ZIP's unzip(1) expects certain options to come before the
1018 * zipfile name, and others to come after - though it does not
1019 * enforce this. For simplicity, we accept *all* options both
1020 * before and after the zipfile name.
1021 */
1022 nopts = getopts(argc, argv);
1023
1024 if (argc <= nopts)
1025 usage();
1026 zipfile = argv[nopts++];
1027
1028 if (strcmp(zipfile, "-") == 0)
1029 zipfile = NULL; /* STDIN */
1030
1031 while (nopts < argc && *argv[nopts] != '-')
1032 add_pattern(&include, argv[nopts++]);
1033
1034 nopts--; /* fake argv[0] */
1035 nopts += getopts(argc - nopts, argv + nopts);
1036
1037 if (n_opt + o_opt + u_opt > 1)
1038 errorx("-n, -o and -u are contradictory");
1039
1040 time(&now);
1041
1042 unzip(zipfile);
1043
1044 exit(0);
1045 }
1046