unzip.c revision 1.23 1 /* $NetBSD: unzip.c,v 1.23 2017/04/20 13:11:35 joerg 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 2017/04/20 13:11:35 joerg 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 mode = archive_entry_mode(e) & 0777;
391 if (mode == 0)
392 mode = 0755;
393
394 /*
395 * Some zipfiles contain directories with weird permissions such
396 * as 0644 or 0444. This can cause strange issues such as being
397 * unable to extract files into the directory we just created, or
398 * the user being unable to remove the directory later without
399 * first manually changing its permissions. Therefore, we whack
400 * the permissions into shape, assuming that the user wants full
401 * access and that anyone who gets read access also gets execute
402 * access.
403 */
404 mode |= 0700;
405 if (mode & 0040)
406 mode |= 0010;
407 if (mode & 0004)
408 mode |= 0001;
409
410 info(" creating: %s/\n", path);
411 make_dir(path, mode);
412 ac(archive_read_data_skip(a));
413 }
414
415 static unsigned char buffer[8192];
416 static char spinner[] = { '|', '/', '-', '\\' };
417
418 static int
419 handle_existing_file(char **path)
420 {
421 size_t alen;
422 ssize_t len;
423 char buf[4];
424
425 for (;;) {
426 fprintf(stderr,
427 "replace %s? [y]es, [n]o, [A]ll, [N]one, [r]ename: ",
428 *path);
429 fgets(buf, 4, stdin);
430 switch (*buf) {
431 case 'A':
432 o_opt = 1;
433 /* FALL THROUGH */
434 case 'y':
435 case 'Y':
436 (void)unlink(*path);
437 return 1;
438 case 'N':
439 n_opt = 1;
440 /* FALL THROUGH */
441 case 'n':
442 return -1;
443 case 'r':
444 case 'R':
445 printf("New name: ");
446 fflush(stdout);
447 free(*path);
448 *path = NULL;
449 alen = 0;
450 len = getline(path, &alen, stdin);
451 if ((*path)[len - 1] == '\n')
452 (*path)[len - 1] = '\0';
453 return 0;
454 default:
455 break;
456 }
457 }
458 }
459
460 /*
461 * Detect binary files by a combination of character white list and
462 * black list. NUL bytes and other control codes without use in text files
463 * result directly in switching the file to binary mode. Otherwise, at least
464 * one white-listed byte has to be found.
465 *
466 * Black-listed: 0..6, 14..25, 28..31
467 * White-listed: 9..10, 13, >= 32
468 *
469 * See the proginfo/txtvsbin.txt in the zip sources for a detailed discussion.
470 */
471 #define BYTE_IS_BINARY(x) ((x) < 32 && (0xf3ffc07fU & (1U << (x))))
472 #define BYTE_IS_TEXT(x) ((x) >= 32 || (0x00002600U & (1U << (x))))
473
474 static int
475 check_binary(const unsigned char *buf, size_t len)
476 {
477 int rv;
478 for (rv = 1; len--; ++buf) {
479 if (BYTE_IS_BINARY(*buf))
480 return 1;
481 if (BYTE_IS_TEXT(*buf))
482 rv = 0;
483 }
484
485 return rv;
486 }
487
488 /*
489 * Extract to a file descriptor
490 */
491 static int
492 extract2fd(struct archive *a, char *pathname, int fd)
493 {
494 int cr, text, warn;
495 ssize_t len;
496 unsigned char *p, *q, *end;
497
498 text = a_opt;
499 warn = 0;
500 cr = 0;
501
502 /* loop over file contents and write to fd */
503 for (int n = 0; ; n++) {
504 if (fd != STDOUT_FILENO)
505 if (tty && (n % 4) == 0)
506 info(" %c\b\b", spinner[(n / 4) % sizeof spinner]);
507
508 len = archive_read_data(a, buffer, sizeof buffer);
509
510 if (len < 0)
511 ac(len);
512
513 /* left over CR from previous buffer */
514 if (a_opt && cr) {
515 if (len == 0 || buffer[0] != '\n')
516 if (write(fd, "\r", 1) != 1)
517 error("write('%s')", pathname);
518 cr = 0;
519 }
520
521 /* EOF */
522 if (len == 0)
523 break;
524 end = buffer + len;
525
526 /*
527 * Detect whether this is a text file. The correct way to
528 * do this is to check the least significant bit of the
529 * "internal file attributes" field of the corresponding
530 * file header in the central directory, but libarchive
531 * does not provide access to this field, so we have to
532 * guess by looking for non-ASCII characters in the
533 * buffer. Hopefully we won't guess wrong. If we do
534 * guess wrong, we print a warning message later.
535 */
536 if (a_opt && n == 0) {
537 if (check_binary(buffer, len))
538 text = 0;
539 }
540
541 /* simple case */
542 if (!a_opt || !text) {
543 if (write(fd, buffer, len) != len)
544 error("write('%s')", pathname);
545 continue;
546 }
547
548 /* hard case: convert \r\n to \n (sigh...) */
549 for (p = buffer; p < end; p = q + 1) {
550 for (q = p; q < end; q++) {
551 if (!warn && BYTE_IS_BINARY(*q)) {
552 warningx("%s may be corrupted due"
553 " to weak text file detection"
554 " heuristic", pathname);
555 warn = 1;
556 }
557 if (q[0] != '\r')
558 continue;
559 if (&q[1] == end) {
560 cr = 1;
561 break;
562 }
563 if (q[1] == '\n')
564 break;
565 }
566 if (write(fd, p, q - p) != q - p)
567 error("write('%s')", pathname);
568 }
569 }
570
571 return text;
572 }
573
574 /*
575 * Extract a regular file.
576 */
577 static void
578 extract_file(struct archive *a, struct archive_entry *e, char **path)
579 {
580 int mode;
581 time_t mtime;
582 struct stat sb;
583 struct timeval tv[2];
584 int fd, check, text;
585 const char *linkname;
586
587 mode = archive_entry_mode(e) & 0777;
588 if (mode == 0)
589 mode = 0644;
590 mtime = archive_entry_mtime(e);
591
592 /* look for existing file of same name */
593 recheck:
594 if (lstat(*path, &sb) == 0) {
595 if (u_opt || f_opt) {
596 /* check if up-to-date */
597 if (S_ISREG(sb.st_mode) && sb.st_mtime >= mtime)
598 return;
599 (void)unlink(*path);
600 } else if (o_opt) {
601 /* overwrite */
602 (void)unlink(*path);
603 } else if (n_opt) {
604 /* do not overwrite */
605 return;
606 } else {
607 check = handle_existing_file(path);
608 if (check == 0)
609 goto recheck;
610 if (check == -1)
611 return; /* do not overwrite */
612 }
613 } else {
614 if (f_opt)
615 return;
616 }
617
618 tv[0].tv_sec = now;
619 tv[0].tv_usec = 0;
620 tv[1].tv_sec = mtime;
621 tv[1].tv_usec = 0;
622
623 /* process symlinks */
624 linkname = archive_entry_symlink(e);
625 if (linkname != NULL) {
626 if (symlink(linkname, *path) == -1)
627 error("symlink('%s', '%s')", linkname, *path);
628 info(" extracting: %s -> %s\n", *path, linkname);
629 if (lchmod(*path, mode) == -1)
630 warning("Cannot set mode for '%s'", *path);
631 if (lutimes(*path, tv) == -1)
632 warning("utimes('%s')", *path);
633 return;
634 }
635
636 /* process hardlinks */
637 linkname = archive_entry_hardlink(e);
638 if (linkname != NULL) {
639 if (link(linkname, *path) == -1)
640 error("link('%s', '%s')", linkname, *path);
641 info(" extracting: %s link to %s\n", *path, linkname);
642 return;
643 }
644
645 if ((fd = open(*path, O_RDWR|O_CREAT|O_TRUNC, mode)) < 0)
646 error("open('%s')", *path);
647
648 info(" extracting: %s", *path);
649
650 text = extract2fd(a, *path, fd);
651
652 if (tty)
653 info(" \b\b");
654 if (text)
655 info(" (text)");
656 info("\n");
657
658 /* set access and modification time */
659 if (futimes(fd, tv) != 0)
660 error("utimes('%s')", *path);
661 if (close(fd) != 0)
662 error("close('%s')", *path);
663 }
664
665 /*
666 * Extract a zipfile entry: first perform some sanity checks to ensure
667 * that it is either a directory or a regular file and that the path is
668 * not absolute and does not try to break out of the current directory;
669 * then call either extract_dir() or extract_file() as appropriate.
670 *
671 * This is complicated a bit by the various ways in which we need to
672 * manipulate the path name. Case conversion (if requested by the -L
673 * option) happens first, but the include / exclude patterns are applied
674 * to the full converted path name, before the directory part of the path
675 * is removed in accordance with the -j option. Sanity checks are
676 * intentionally done earlier than they need to be, so the user will get a
677 * warning about insecure paths even for files or directories which
678 * wouldn't be extracted anyway.
679 */
680 static void
681 extract(struct archive *a, struct archive_entry *e)
682 {
683 char *pathname, *realpathname;
684 mode_t filetype;
685 char *p, *q;
686
687 pathname = pathdup(archive_entry_pathname(e));
688 filetype = archive_entry_filetype(e);
689
690 /* sanity checks */
691 if (pathname[0] == '/' ||
692 strncmp(pathname, "../", 3) == 0 ||
693 strstr(pathname, "/../") != NULL) {
694 warningx("skipping insecure entry '%s'", pathname);
695 ac(archive_read_data_skip(a));
696 free(pathname);
697 return;
698 }
699
700 /* I don't think this can happen in a zipfile.. */
701 if (!S_ISDIR(filetype) && !S_ISREG(filetype) && !S_ISLNK(filetype)) {
702 warningx("skipping non-regular entry '%s'", pathname);
703 ac(archive_read_data_skip(a));
704 free(pathname);
705 return;
706 }
707
708 /* skip directories in -j case */
709 if (S_ISDIR(filetype) && j_opt) {
710 ac(archive_read_data_skip(a));
711 free(pathname);
712 return;
713 }
714
715 /* apply include / exclude patterns */
716 if (!accept_pathname(pathname)) {
717 ac(archive_read_data_skip(a));
718 free(pathname);
719 return;
720 }
721
722 /* apply -j and -d */
723 if (j_opt) {
724 for (p = q = pathname; *p; ++p)
725 if (*p == '/')
726 q = p + 1;
727 realpathname = pathcat(d_arg, q);
728 } else {
729 realpathname = pathcat(d_arg, pathname);
730 }
731
732 /* ensure that parent directory exists */
733 make_parent(realpathname);
734
735 if (S_ISDIR(filetype))
736 extract_dir(a, e, realpathname);
737 else
738 extract_file(a, e, &realpathname);
739
740 free(realpathname);
741 free(pathname);
742 }
743
744 static void
745 extract_stdout(struct archive *a, struct archive_entry *e)
746 {
747 char *pathname;
748 mode_t filetype;
749
750 pathname = pathdup(archive_entry_pathname(e));
751 filetype = archive_entry_filetype(e);
752
753 /* I don't think this can happen in a zipfile.. */
754 if (!S_ISDIR(filetype) && !S_ISREG(filetype) && !S_ISLNK(filetype)) {
755 warningx("skipping non-regular entry '%s'", pathname);
756 ac(archive_read_data_skip(a));
757 free(pathname);
758 return;
759 }
760
761 /* skip directories in -j case */
762 if (S_ISDIR(filetype)) {
763 ac(archive_read_data_skip(a));
764 free(pathname);
765 return;
766 }
767
768 /* apply include / exclude patterns */
769 if (!accept_pathname(pathname)) {
770 ac(archive_read_data_skip(a));
771 free(pathname);
772 return;
773 }
774
775 if (c_opt)
776 info("x %s\n", pathname);
777
778 (void)extract2fd(a, pathname, STDOUT_FILENO);
779
780 free(pathname);
781 }
782
783 /*
784 * Print the name of an entry to stdout.
785 */
786 static void
787 list(struct archive *a, struct archive_entry *e)
788 {
789 char buf[20];
790 time_t mtime;
791 struct tm *tm;
792
793 mtime = archive_entry_mtime(e);
794 tm = localtime(&mtime);
795 if (*y_str)
796 strftime(buf, sizeof(buf), "%m-%d-%G %R", tm);
797 else
798 strftime(buf, sizeof(buf), "%m-%d-%g %R", tm);
799
800 if (v_opt == 1) {
801 printf(" %8ju %s %s\n",
802 (uintmax_t)archive_entry_size(e),
803 buf, archive_entry_pathname(e));
804 } else if (v_opt == 2) {
805 printf("%8ju Stored %7ju 0%% %s %08x %s\n",
806 (uintmax_t)archive_entry_size(e),
807 (uintmax_t)archive_entry_size(e),
808 buf,
809 0U,
810 archive_entry_pathname(e));
811 }
812 ac(archive_read_data_skip(a));
813 }
814
815 /*
816 * Extract to memory to check CRC
817 */
818 static int
819 test(struct archive *a, struct archive_entry *e)
820 {
821 ssize_t len;
822 int error_count;
823
824 error_count = 0;
825 if (S_ISDIR(archive_entry_filetype(e)))
826 return 0;
827
828 info(" testing: %s\t", archive_entry_pathname(e));
829 while ((len = archive_read_data(a, buffer, sizeof buffer)) > 0)
830 /* nothing */;
831 if (len < 0) {
832 info(" %s\n", archive_error_string(a));
833 ++error_count;
834 } else {
835 info(" OK\n");
836 }
837
838 /* shouldn't be necessary, but it doesn't hurt */
839 ac(archive_read_data_skip(a));
840
841 return error_count;
842 }
843
844 /*
845 * Main loop: open the zipfile, iterate over its contents and decide what
846 * to do with each entry.
847 */
848 static void
849 unzip(const char *fn)
850 {
851 struct archive *a;
852 struct archive_entry *e;
853 int ret;
854 uintmax_t total_size, file_count, error_count;
855
856 if ((a = archive_read_new()) == NULL)
857 error("archive_read_new failed");
858
859 ac(archive_read_support_format_zip(a));
860 ac(archive_read_open_filename(a, fn, 8192));
861
862 if (!q_opt && !p_opt)
863 printf("Archive: %s\n", fn);
864
865 if (v_opt == 1) {
866 printf(" Length %sDate Time Name\n", y_str);
867 printf(" -------- %s---- ---- ----\n", y_str);
868 } else if (v_opt == 2) {
869 printf(" Length Method Size Ratio %sDate Time CRC-32 Name\n", y_str);
870 printf("-------- ------ ------- ----- %s---- ---- ------ ----\n", y_str);
871 }
872
873 total_size = 0;
874 file_count = 0;
875 error_count = 0;
876 for (;;) {
877 ret = archive_read_next_header(a, &e);
878 if (ret == ARCHIVE_EOF)
879 break;
880 ac(ret);
881 if (t_opt)
882 error_count += test(a, e);
883 else if (v_opt)
884 list(a, e);
885 else if (p_opt || c_opt)
886 extract_stdout(a, e);
887 else
888 extract(a, e);
889
890 total_size += archive_entry_size(e);
891 ++file_count;
892 }
893
894 if (v_opt == 1) {
895 printf(" -------- %s-------\n", y_str);
896 printf(" %8ju %s%ju file%s\n",
897 total_size, y_str, file_count, file_count != 1 ? "s" : "");
898 } else if (v_opt == 2) {
899 printf("-------- ------- --- %s-------\n", y_str);
900 printf("%8ju %7ju 0%% %s%ju file%s\n",
901 total_size, total_size, y_str, file_count,
902 file_count != 1 ? "s" : "");
903 }
904
905 ac(archive_read_free(a));
906
907 if (t_opt) {
908 if (error_count > 0) {
909 errorx("%ju checksum error(s) found.", error_count);
910 }
911 else {
912 printf("No errors detected in compressed data of %s.\n",
913 fn);
914 }
915 }
916 }
917
918 static void __dead
919 usage(void)
920 {
921
922 fprintf(stderr, "Usage: %s [-aCcfjLlnopqtuvy] [-d dir] [-x pattern] "
923 "zipfile\n", getprogname());
924 exit(1);
925 }
926
927 static int
928 getopts(int argc, char *argv[])
929 {
930 int opt;
931
932 optreset = optind = 1;
933 while ((opt = getopt(argc, argv, "aCcd:fjLlnopqtuvyx:")) != -1)
934 switch (opt) {
935 case 'a':
936 a_opt = 1;
937 break;
938 case 'C':
939 C_opt = 1;
940 break;
941 case 'c':
942 c_opt = 1;
943 break;
944 case 'd':
945 d_arg = optarg;
946 break;
947 case 'f':
948 f_opt = 1;
949 break;
950 case 'j':
951 j_opt = 1;
952 break;
953 case 'L':
954 L_opt = 1;
955 break;
956 case 'l':
957 if (v_opt == 0)
958 v_opt = 1;
959 break;
960 case 'n':
961 n_opt = 1;
962 break;
963 case 'o':
964 o_opt = 1;
965 q_opt = 1;
966 break;
967 case 'p':
968 p_opt = 1;
969 break;
970 case 'q':
971 q_opt = 1;
972 break;
973 case 't':
974 t_opt = 1;
975 break;
976 case 'u':
977 u_opt = 1;
978 break;
979 case 'v':
980 v_opt = 2;
981 break;
982 case 'x':
983 add_pattern(&exclude, optarg);
984 break;
985 case 'y':
986 y_str = " ";
987 break;
988 default:
989 usage();
990 }
991
992 return (optind);
993 }
994
995 int
996 main(int argc, char *argv[])
997 {
998 const char *zipfile;
999 int nopts;
1000
1001 if (isatty(STDOUT_FILENO))
1002 tty = 1;
1003
1004 if (getenv("UNZIP_DEBUG") != NULL)
1005 unzip_debug = 1;
1006 for (int i = 0; i < argc; ++i)
1007 debug("%s%c", argv[i], (i < argc - 1) ? ' ' : '\n');
1008
1009 /*
1010 * Info-ZIP's unzip(1) expects certain options to come before the
1011 * zipfile name, and others to come after - though it does not
1012 * enforce this. For simplicity, we accept *all* options both
1013 * before and after the zipfile name.
1014 */
1015 nopts = getopts(argc, argv);
1016
1017 if (argc <= nopts)
1018 usage();
1019 zipfile = argv[nopts++];
1020
1021 if (strcmp(zipfile, "-") == 0)
1022 zipfile = NULL; /* STDIN */
1023
1024 while (nopts < argc && *argv[nopts] != '-')
1025 add_pattern(&include, argv[nopts++]);
1026
1027 nopts--; /* fake argv[0] */
1028 nopts += getopts(argc - nopts, argv + nopts);
1029
1030 if (n_opt + o_opt + u_opt > 1)
1031 errorx("-n, -o and -u are contradictory");
1032
1033 time(&now);
1034
1035 unzip(zipfile);
1036
1037 exit(0);
1038 }
1039