Deleted Added
full compact
unzip.c (236226) unzip.c (248612)
1/*-
2 * Copyright (c) 2009 Joerg Sonnenberger <joerg@NetBSD.org>
3 * Copyright (c) 2007-2008 Dag-Erling Sm��rgrav
4 * All rights reserved.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer
11 * in this position and unchanged.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 *
1/*-
2 * Copyright (c) 2009 Joerg Sonnenberger <joerg@NetBSD.org>
3 * Copyright (c) 2007-2008 Dag-Erling Sm��rgrav
4 * All rights reserved.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer
11 * in this position and unchanged.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 *
28 * $FreeBSD: head/usr.bin/unzip/unzip.c 236226 2012-05-29 09:11:19Z des $
28 * $FreeBSD: head/usr.bin/unzip/unzip.c 248612 2013-03-22 10:17:42Z mm $
29 *
30 * This file would be much shorter if we didn't care about command-line
31 * compatibility with Info-ZIP's UnZip, which requires us to duplicate
32 * parts of libarchive in order to gain more detailed control of its
33 * behaviour for the purpose of implementing the -n, -o, -L and -a
34 * options.
35 */
36
37#include <sys/queue.h>
38#include <sys/stat.h>
39
40#include <ctype.h>
41#include <errno.h>
42#include <fcntl.h>
43#include <fnmatch.h>
44#include <stdarg.h>
45#include <stdio.h>
46#include <stdlib.h>
47#include <string.h>
48#include <unistd.h>
49
50#include <archive.h>
51#include <archive_entry.h>
52
53/* command-line options */
54static int a_opt; /* convert EOL */
55static int C_opt; /* match case-insensitively */
56static int c_opt; /* extract to stdout */
57static const char *d_arg; /* directory */
58static int f_opt; /* update existing files only */
59static int j_opt; /* junk directories */
60static int L_opt; /* lowercase names */
61static int n_opt; /* never overwrite */
62static int o_opt; /* always overwrite */
63static int p_opt; /* extract to stdout, quiet */
64static int q_opt; /* quiet */
65static int t_opt; /* test */
66static int u_opt; /* update */
67static int v_opt; /* verbose/list */
68static int Z1_opt; /* zipinfo mode list files only */
69
70/* time when unzip started */
71static time_t now;
72
73/* debug flag */
74static int unzip_debug;
75
76/* zipinfo mode */
77static int zipinfo_mode;
78
79/* running on tty? */
80static int tty;
81
82/* convenience macro */
83/* XXX should differentiate between ARCHIVE_{WARN,FAIL,RETRY} */
84#define ac(call) \
85 do { \
86 int acret = (call); \
87 if (acret != ARCHIVE_OK) \
88 errorx("%s", archive_error_string(a)); \
89 } while (0)
90
91/*
92 * Indicates that last info() did not end with EOL. This helps error() et
93 * al. avoid printing an error message on the same line as an incomplete
94 * informational message.
95 */
96static int noeol;
97
98/* fatal error message + errno */
99static void
100error(const char *fmt, ...)
101{
102 va_list ap;
103
104 if (noeol)
105 fprintf(stdout, "\n");
106 fflush(stdout);
107 fprintf(stderr, "unzip: ");
108 va_start(ap, fmt);
109 vfprintf(stderr, fmt, ap);
110 va_end(ap);
111 fprintf(stderr, ": %s\n", strerror(errno));
112 exit(1);
113}
114
115/* fatal error message, no errno */
116static void
117errorx(const char *fmt, ...)
118{
119 va_list ap;
120
121 if (noeol)
122 fprintf(stdout, "\n");
123 fflush(stdout);
124 fprintf(stderr, "unzip: ");
125 va_start(ap, fmt);
126 vfprintf(stderr, fmt, ap);
127 va_end(ap);
128 fprintf(stderr, "\n");
129 exit(1);
130}
131
132#if 0
133/* non-fatal error message + errno */
134static void
135warning(const char *fmt, ...)
136{
137 va_list ap;
138
139 if (noeol)
140 fprintf(stdout, "\n");
141 fflush(stdout);
142 fprintf(stderr, "unzip: ");
143 va_start(ap, fmt);
144 vfprintf(stderr, fmt, ap);
145 va_end(ap);
146 fprintf(stderr, ": %s\n", strerror(errno));
147}
148#endif
149
150/* non-fatal error message, no errno */
151static void
152warningx(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) */
167static void
168info(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) */
186static void
187debug(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 */
205static char *
206pathdup(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 */
230static char *
231pathcat(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 */
254struct pattern {
255 STAILQ_ENTRY(pattern) link;
256 char pattern[];
257};
258
259STAILQ_HEAD(pattern_list, pattern);
260static struct pattern_list include = STAILQ_HEAD_INITIALIZER(include);
261static struct pattern_list exclude = STAILQ_HEAD_INITIALIZER(exclude);
262
263/*
264 * Add an entry to a pattern list
265 */
266static void
267add_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 */
285static int
286match_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 */
301static int
302accept_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 */
316static void
317make_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 */
349static void
350make_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 */
385static void
386extract_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("d %s\n", path);
411 make_dir(path, mode);
412 ac(archive_read_data_skip(a));
413}
414
415static unsigned char buffer[8192];
416static char spinner[] = { '|', '/', '-', '\\' };
417
418static int
419handle_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 if (fgets(buf, sizeof(buf), stdin) == NULL) {
430 clearerr(stdin);
431 printf("NULL\n(EOF or read error, "
432 "treating as \"[N]one\"...)\n");
433 n_opt = 1;
434 return -1;
435 }
436 switch (*buf) {
437 case 'A':
438 o_opt = 1;
439 /* FALLTHROUGH */
440 case 'y':
441 case 'Y':
442 (void)unlink(*path);
443 return 1;
444 case 'N':
445 n_opt = 1;
446 /* FALLTHROUGH */
447 case 'n':
448 return -1;
449 case 'r':
450 case 'R':
451 printf("New name: ");
452 fflush(stdout);
453 free(*path);
454 *path = NULL;
455 alen = 0;
456 len = getdelim(path, &alen, '\n', stdin);
457 if ((*path)[len - 1] == '\n')
458 (*path)[len - 1] = '\0';
459 return 0;
460 default:
461 break;
462 }
463 }
464}
465
466/*
467 * Extract a regular file.
468 */
469static void
470extract_file(struct archive *a, struct archive_entry *e, char **path)
471{
472 int mode;
473 time_t mtime;
474 struct stat sb;
475 struct timeval tv[2];
476 int cr, fd, text, warn, check;
477 ssize_t len;
478 unsigned char *p, *q, *end;
479
480 mode = archive_entry_mode(e) & 0777;
481 if (mode == 0)
482 mode = 0644;
483 mtime = archive_entry_mtime(e);
484
485 /* look for existing file of same name */
486recheck:
487 if (lstat(*path, &sb) == 0) {
488 if (u_opt || f_opt) {
489 /* check if up-to-date */
490 if (S_ISREG(sb.st_mode) && sb.st_mtime >= mtime)
491 return;
492 (void)unlink(*path);
493 } else if (o_opt) {
494 /* overwrite */
495 (void)unlink(*path);
496 } else if (n_opt) {
497 /* do not overwrite */
498 return;
499 } else {
500 check = handle_existing_file(path);
501 if (check == 0)
502 goto recheck;
503 if (check == -1)
504 return; /* do not overwrite */
505 }
506 } else {
507 if (f_opt)
508 return;
509 }
510
511 if ((fd = open(*path, O_RDWR|O_CREAT|O_TRUNC, mode)) < 0)
512 error("open('%s')", *path);
513
514 /* loop over file contents and write to disk */
515 info(" extracting: %s", *path);
516 text = a_opt;
517 warn = 0;
518 cr = 0;
519 for (int n = 0; ; n++) {
520 if (tty && (n % 4) == 0)
521 info(" %c\b\b", spinner[(n / 4) % sizeof spinner]);
522
523 len = archive_read_data(a, buffer, sizeof buffer);
524
525 if (len < 0)
526 ac(len);
527
528 /* left over CR from previous buffer */
529 if (a_opt && cr) {
530 if (len == 0 || buffer[0] != '\n')
531 if (write(fd, "\r", 1) != 1)
532 error("write('%s')", *path);
533 cr = 0;
534 }
535
536 /* EOF */
537 if (len == 0)
538 break;
539 end = buffer + len;
540
541 /*
542 * Detect whether this is a text file. The correct way to
543 * do this is to check the least significant bit of the
544 * "internal file attributes" field of the corresponding
545 * file header in the central directory, but libarchive
546 * does not read the central directory, so we have to
547 * guess by looking for non-ASCII characters in the
548 * buffer. Hopefully we won't guess wrong. If we do
549 * guess wrong, we print a warning message later.
550 */
551 if (a_opt && n == 0) {
552 for (p = buffer; p < end; ++p) {
553 if (!isascii((unsigned char)*p)) {
554 text = 0;
555 break;
556 }
557 }
558 }
559
560 /* simple case */
561 if (!a_opt || !text) {
562 if (write(fd, buffer, len) != len)
563 error("write('%s')", *path);
564 continue;
565 }
566
567 /* hard case: convert \r\n to \n (sigh...) */
568 for (p = buffer; p < end; p = q + 1) {
569 for (q = p; q < end; q++) {
570 if (!warn && !isascii(*q)) {
571 warningx("%s may be corrupted due"
572 " to weak text file detection"
573 " heuristic", *path);
574 warn = 1;
575 }
576 if (q[0] != '\r')
577 continue;
578 if (&q[1] == end) {
579 cr = 1;
580 break;
581 }
582 if (q[1] == '\n')
583 break;
584 }
585 if (write(fd, p, q - p) != q - p)
586 error("write('%s')", *path);
587 }
588 }
589 if (tty)
590 info(" \b\b");
591 if (text)
592 info(" (text)");
593 info("\n");
594
595 /* set access and modification time */
596 tv[0].tv_sec = now;
597 tv[0].tv_usec = 0;
598 tv[1].tv_sec = mtime;
599 tv[1].tv_usec = 0;
600 if (futimes(fd, tv) != 0)
601 error("utimes('%s')", *path);
602 if (close(fd) != 0)
603 error("close('%s')", *path);
604}
605
606/*
607 * Extract a zipfile entry: first perform some sanity checks to ensure
608 * that it is either a directory or a regular file and that the path is
609 * not absolute and does not try to break out of the current directory;
610 * then call either extract_dir() or extract_file() as appropriate.
611 *
612 * This is complicated a bit by the various ways in which we need to
613 * manipulate the path name. Case conversion (if requested by the -L
614 * option) happens first, but the include / exclude patterns are applied
615 * to the full converted path name, before the directory part of the path
616 * is removed in accordance with the -j option. Sanity checks are
617 * intentionally done earlier than they need to be, so the user will get a
618 * warning about insecure paths even for files or directories which
619 * wouldn't be extracted anyway.
620 */
621static void
622extract(struct archive *a, struct archive_entry *e)
623{
624 char *pathname, *realpathname;
625 mode_t filetype;
626 char *p, *q;
627
628 pathname = pathdup(archive_entry_pathname(e));
629 filetype = archive_entry_filetype(e);
630
631 /* sanity checks */
632 if (pathname[0] == '/' ||
633 strncmp(pathname, "../", 3) == 0 ||
634 strstr(pathname, "/../") != NULL) {
635 warningx("skipping insecure entry '%s'", pathname);
636 ac(archive_read_data_skip(a));
637 free(pathname);
638 return;
639 }
640
641 /* I don't think this can happen in a zipfile.. */
642 if (!S_ISDIR(filetype) && !S_ISREG(filetype)) {
643 warningx("skipping non-regular entry '%s'", pathname);
644 ac(archive_read_data_skip(a));
645 free(pathname);
646 return;
647 }
648
649 /* skip directories in -j case */
650 if (S_ISDIR(filetype) && j_opt) {
651 ac(archive_read_data_skip(a));
652 free(pathname);
653 return;
654 }
655
656 /* apply include / exclude patterns */
657 if (!accept_pathname(pathname)) {
658 ac(archive_read_data_skip(a));
659 free(pathname);
660 return;
661 }
662
663 /* apply -j and -d */
664 if (j_opt) {
665 for (p = q = pathname; *p; ++p)
666 if (*p == '/')
667 q = p + 1;
668 realpathname = pathcat(d_arg, q);
669 } else {
670 realpathname = pathcat(d_arg, pathname);
671 }
672
673 /* ensure that parent directory exists */
674 make_parent(realpathname);
675
676 if (S_ISDIR(filetype))
677 extract_dir(a, e, realpathname);
678 else
679 extract_file(a, e, &realpathname);
680
681 free(realpathname);
682 free(pathname);
683}
684
685static void
686extract_stdout(struct archive *a, struct archive_entry *e)
687{
688 char *pathname;
689 mode_t filetype;
690 int cr, text, warn;
691 ssize_t len;
692 unsigned char *p, *q, *end;
693
694 pathname = pathdup(archive_entry_pathname(e));
695 filetype = archive_entry_filetype(e);
696
697 /* I don't think this can happen in a zipfile.. */
698 if (!S_ISDIR(filetype) && !S_ISREG(filetype)) {
699 warningx("skipping non-regular entry '%s'", pathname);
700 ac(archive_read_data_skip(a));
701 free(pathname);
702 return;
703 }
704
705 /* skip directories in -j case */
706 if (S_ISDIR(filetype)) {
707 ac(archive_read_data_skip(a));
708 free(pathname);
709 return;
710 }
711
712 /* apply include / exclude patterns */
713 if (!accept_pathname(pathname)) {
714 ac(archive_read_data_skip(a));
715 free(pathname);
716 return;
717 }
718
719 if (c_opt)
720 info("x %s\n", pathname);
721
722 text = a_opt;
723 warn = 0;
724 cr = 0;
725 for (int n = 0; ; n++) {
726 len = archive_read_data(a, buffer, sizeof buffer);
727
728 if (len < 0)
729 ac(len);
730
731 /* left over CR from previous buffer */
732 if (a_opt && cr) {
733 if (len == 0 || buffer[0] != '\n') {
734 if (fwrite("\r", 1, 1, stderr) != 1)
735 error("write('%s')", pathname);
736 }
737 cr = 0;
738 }
739
740 /* EOF */
741 if (len == 0)
742 break;
743 end = buffer + len;
744
745 /*
746 * Detect whether this is a text file. The correct way to
747 * do this is to check the least significant bit of the
748 * "internal file attributes" field of the corresponding
749 * file header in the central directory, but libarchive
750 * does not read the central directory, so we have to
751 * guess by looking for non-ASCII characters in the
752 * buffer. Hopefully we won't guess wrong. If we do
753 * guess wrong, we print a warning message later.
754 */
755 if (a_opt && n == 0) {
756 for (p = buffer; p < end; ++p) {
757 if (!isascii((unsigned char)*p)) {
758 text = 0;
759 break;
760 }
761 }
762 }
763
764 /* simple case */
765 if (!a_opt || !text) {
766 if (fwrite(buffer, 1, len, stdout) != (size_t)len)
767 error("write('%s')", pathname);
768 continue;
769 }
770
771 /* hard case: convert \r\n to \n (sigh...) */
772 for (p = buffer; p < end; p = q + 1) {
773 for (q = p; q < end; q++) {
774 if (!warn && !isascii(*q)) {
775 warningx("%s may be corrupted due"
776 " to weak text file detection"
777 " heuristic", pathname);
778 warn = 1;
779 }
780 if (q[0] != '\r')
781 continue;
782 if (&q[1] == end) {
783 cr = 1;
784 break;
785 }
786 if (q[1] == '\n')
787 break;
788 }
789 if (fwrite(p, 1, q - p, stdout) != (size_t)(q - p))
790 error("write('%s')", pathname);
791 }
792 }
793
794 free(pathname);
795}
796
797/*
798 * Print the name of an entry to stdout.
799 */
800static void
801list(struct archive *a, struct archive_entry *e)
802{
803 char buf[20];
804 time_t mtime;
805
806 mtime = archive_entry_mtime(e);
807 strftime(buf, sizeof(buf), "%m-%d-%g %R", localtime(&mtime));
808
809 if (!zipinfo_mode) {
810 if (v_opt == 1) {
811 printf(" %8ju %s %s\n",
812 (uintmax_t)archive_entry_size(e),
813 buf, archive_entry_pathname(e));
814 } else if (v_opt == 2) {
815 printf("%8ju Stored %7ju 0%% %s %08x %s\n",
816 (uintmax_t)archive_entry_size(e),
817 (uintmax_t)archive_entry_size(e),
818 buf,
819 0U,
820 archive_entry_pathname(e));
821 }
822 } else {
823 if (Z1_opt)
824 printf("%s\n",archive_entry_pathname(e));
825 }
826 ac(archive_read_data_skip(a));
827}
828
829/*
830 * Extract to memory to check CRC
831 */
832static int
833test(struct archive *a, struct archive_entry *e)
834{
835 ssize_t len;
836 int error_count;
837
838 error_count = 0;
839 if (S_ISDIR(archive_entry_filetype(e)))
840 return 0;
841
842 info(" testing: %s\t", archive_entry_pathname(e));
843 while ((len = archive_read_data(a, buffer, sizeof buffer)) > 0)
844 /* nothing */;
845 if (len < 0) {
846 info(" %s\n", archive_error_string(a));
847 ++error_count;
848 } else {
849 info(" OK\n");
850 }
851
852 /* shouldn't be necessary, but it doesn't hurt */
853 ac(archive_read_data_skip(a));
854
855 return error_count;
856}
857
858
859/*
860 * Main loop: open the zipfile, iterate over its contents and decide what
861 * to do with each entry.
862 */
863static void
864unzip(const char *fn)
865{
866 struct archive *a;
867 struct archive_entry *e;
868 int ret;
869 uintmax_t total_size, file_count, error_count;
870
871 if ((a = archive_read_new()) == NULL)
872 error("archive_read_new failed");
873
874 ac(archive_read_support_format_zip(a));
875 ac(archive_read_open_filename(a, fn, 8192));
876
877 if (!zipinfo_mode) {
878 if (!p_opt && !q_opt)
879 printf("Archive: %s\n", fn);
880 if (v_opt == 1) {
881 printf(" Length Date Time Name\n");
882 printf(" -------- ---- ---- ----\n");
883 } else if (v_opt == 2) {
884 printf(" Length Method Size Ratio Date Time CRC-32 Name\n");
885 printf("-------- ------ ------- ----- ---- ---- ------ ----\n");
886 }
887 }
888
889 total_size = 0;
890 file_count = 0;
891 error_count = 0;
892 for (;;) {
893 ret = archive_read_next_header(a, &e);
894 if (ret == ARCHIVE_EOF)
895 break;
896 ac(ret);
897 if (!zipinfo_mode) {
898 if (t_opt)
899 error_count += test(a, e);
900 else if (v_opt)
901 list(a, e);
902 else if (p_opt || c_opt)
903 extract_stdout(a, e);
904 else
905 extract(a, e);
906 } else {
907 if (Z1_opt)
908 list(a, e);
909 }
910
911 total_size += archive_entry_size(e);
912 ++file_count;
913 }
914
915 if (zipinfo_mode) {
916 if (v_opt == 1) {
917 printf(" -------- -------\n");
918 printf(" %8ju %ju file%s\n",
919 total_size, file_count, file_count != 1 ? "s" : "");
920 } else if (v_opt == 2) {
921 printf("-------- ------- --- -------\n");
922 printf("%8ju %7ju 0%% %ju file%s\n",
923 total_size, total_size, file_count,
924 file_count != 1 ? "s" : "");
925 }
926 }
927
928 ac(archive_read_close(a));
29 *
30 * This file would be much shorter if we didn't care about command-line
31 * compatibility with Info-ZIP's UnZip, which requires us to duplicate
32 * parts of libarchive in order to gain more detailed control of its
33 * behaviour for the purpose of implementing the -n, -o, -L and -a
34 * options.
35 */
36
37#include <sys/queue.h>
38#include <sys/stat.h>
39
40#include <ctype.h>
41#include <errno.h>
42#include <fcntl.h>
43#include <fnmatch.h>
44#include <stdarg.h>
45#include <stdio.h>
46#include <stdlib.h>
47#include <string.h>
48#include <unistd.h>
49
50#include <archive.h>
51#include <archive_entry.h>
52
53/* command-line options */
54static int a_opt; /* convert EOL */
55static int C_opt; /* match case-insensitively */
56static int c_opt; /* extract to stdout */
57static const char *d_arg; /* directory */
58static int f_opt; /* update existing files only */
59static int j_opt; /* junk directories */
60static int L_opt; /* lowercase names */
61static int n_opt; /* never overwrite */
62static int o_opt; /* always overwrite */
63static int p_opt; /* extract to stdout, quiet */
64static int q_opt; /* quiet */
65static int t_opt; /* test */
66static int u_opt; /* update */
67static int v_opt; /* verbose/list */
68static int Z1_opt; /* zipinfo mode list files only */
69
70/* time when unzip started */
71static time_t now;
72
73/* debug flag */
74static int unzip_debug;
75
76/* zipinfo mode */
77static int zipinfo_mode;
78
79/* running on tty? */
80static int tty;
81
82/* convenience macro */
83/* XXX should differentiate between ARCHIVE_{WARN,FAIL,RETRY} */
84#define ac(call) \
85 do { \
86 int acret = (call); \
87 if (acret != ARCHIVE_OK) \
88 errorx("%s", archive_error_string(a)); \
89 } while (0)
90
91/*
92 * Indicates that last info() did not end with EOL. This helps error() et
93 * al. avoid printing an error message on the same line as an incomplete
94 * informational message.
95 */
96static int noeol;
97
98/* fatal error message + errno */
99static void
100error(const char *fmt, ...)
101{
102 va_list ap;
103
104 if (noeol)
105 fprintf(stdout, "\n");
106 fflush(stdout);
107 fprintf(stderr, "unzip: ");
108 va_start(ap, fmt);
109 vfprintf(stderr, fmt, ap);
110 va_end(ap);
111 fprintf(stderr, ": %s\n", strerror(errno));
112 exit(1);
113}
114
115/* fatal error message, no errno */
116static void
117errorx(const char *fmt, ...)
118{
119 va_list ap;
120
121 if (noeol)
122 fprintf(stdout, "\n");
123 fflush(stdout);
124 fprintf(stderr, "unzip: ");
125 va_start(ap, fmt);
126 vfprintf(stderr, fmt, ap);
127 va_end(ap);
128 fprintf(stderr, "\n");
129 exit(1);
130}
131
132#if 0
133/* non-fatal error message + errno */
134static void
135warning(const char *fmt, ...)
136{
137 va_list ap;
138
139 if (noeol)
140 fprintf(stdout, "\n");
141 fflush(stdout);
142 fprintf(stderr, "unzip: ");
143 va_start(ap, fmt);
144 vfprintf(stderr, fmt, ap);
145 va_end(ap);
146 fprintf(stderr, ": %s\n", strerror(errno));
147}
148#endif
149
150/* non-fatal error message, no errno */
151static void
152warningx(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) */
167static void
168info(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) */
186static void
187debug(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 */
205static char *
206pathdup(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 */
230static char *
231pathcat(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 */
254struct pattern {
255 STAILQ_ENTRY(pattern) link;
256 char pattern[];
257};
258
259STAILQ_HEAD(pattern_list, pattern);
260static struct pattern_list include = STAILQ_HEAD_INITIALIZER(include);
261static struct pattern_list exclude = STAILQ_HEAD_INITIALIZER(exclude);
262
263/*
264 * Add an entry to a pattern list
265 */
266static void
267add_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 */
285static int
286match_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 */
301static int
302accept_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 */
316static void
317make_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 */
349static void
350make_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 */
385static void
386extract_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("d %s\n", path);
411 make_dir(path, mode);
412 ac(archive_read_data_skip(a));
413}
414
415static unsigned char buffer[8192];
416static char spinner[] = { '|', '/', '-', '\\' };
417
418static int
419handle_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 if (fgets(buf, sizeof(buf), stdin) == NULL) {
430 clearerr(stdin);
431 printf("NULL\n(EOF or read error, "
432 "treating as \"[N]one\"...)\n");
433 n_opt = 1;
434 return -1;
435 }
436 switch (*buf) {
437 case 'A':
438 o_opt = 1;
439 /* FALLTHROUGH */
440 case 'y':
441 case 'Y':
442 (void)unlink(*path);
443 return 1;
444 case 'N':
445 n_opt = 1;
446 /* FALLTHROUGH */
447 case 'n':
448 return -1;
449 case 'r':
450 case 'R':
451 printf("New name: ");
452 fflush(stdout);
453 free(*path);
454 *path = NULL;
455 alen = 0;
456 len = getdelim(path, &alen, '\n', stdin);
457 if ((*path)[len - 1] == '\n')
458 (*path)[len - 1] = '\0';
459 return 0;
460 default:
461 break;
462 }
463 }
464}
465
466/*
467 * Extract a regular file.
468 */
469static void
470extract_file(struct archive *a, struct archive_entry *e, char **path)
471{
472 int mode;
473 time_t mtime;
474 struct stat sb;
475 struct timeval tv[2];
476 int cr, fd, text, warn, check;
477 ssize_t len;
478 unsigned char *p, *q, *end;
479
480 mode = archive_entry_mode(e) & 0777;
481 if (mode == 0)
482 mode = 0644;
483 mtime = archive_entry_mtime(e);
484
485 /* look for existing file of same name */
486recheck:
487 if (lstat(*path, &sb) == 0) {
488 if (u_opt || f_opt) {
489 /* check if up-to-date */
490 if (S_ISREG(sb.st_mode) && sb.st_mtime >= mtime)
491 return;
492 (void)unlink(*path);
493 } else if (o_opt) {
494 /* overwrite */
495 (void)unlink(*path);
496 } else if (n_opt) {
497 /* do not overwrite */
498 return;
499 } else {
500 check = handle_existing_file(path);
501 if (check == 0)
502 goto recheck;
503 if (check == -1)
504 return; /* do not overwrite */
505 }
506 } else {
507 if (f_opt)
508 return;
509 }
510
511 if ((fd = open(*path, O_RDWR|O_CREAT|O_TRUNC, mode)) < 0)
512 error("open('%s')", *path);
513
514 /* loop over file contents and write to disk */
515 info(" extracting: %s", *path);
516 text = a_opt;
517 warn = 0;
518 cr = 0;
519 for (int n = 0; ; n++) {
520 if (tty && (n % 4) == 0)
521 info(" %c\b\b", spinner[(n / 4) % sizeof spinner]);
522
523 len = archive_read_data(a, buffer, sizeof buffer);
524
525 if (len < 0)
526 ac(len);
527
528 /* left over CR from previous buffer */
529 if (a_opt && cr) {
530 if (len == 0 || buffer[0] != '\n')
531 if (write(fd, "\r", 1) != 1)
532 error("write('%s')", *path);
533 cr = 0;
534 }
535
536 /* EOF */
537 if (len == 0)
538 break;
539 end = buffer + len;
540
541 /*
542 * Detect whether this is a text file. The correct way to
543 * do this is to check the least significant bit of the
544 * "internal file attributes" field of the corresponding
545 * file header in the central directory, but libarchive
546 * does not read the central directory, so we have to
547 * guess by looking for non-ASCII characters in the
548 * buffer. Hopefully we won't guess wrong. If we do
549 * guess wrong, we print a warning message later.
550 */
551 if (a_opt && n == 0) {
552 for (p = buffer; p < end; ++p) {
553 if (!isascii((unsigned char)*p)) {
554 text = 0;
555 break;
556 }
557 }
558 }
559
560 /* simple case */
561 if (!a_opt || !text) {
562 if (write(fd, buffer, len) != len)
563 error("write('%s')", *path);
564 continue;
565 }
566
567 /* hard case: convert \r\n to \n (sigh...) */
568 for (p = buffer; p < end; p = q + 1) {
569 for (q = p; q < end; q++) {
570 if (!warn && !isascii(*q)) {
571 warningx("%s may be corrupted due"
572 " to weak text file detection"
573 " heuristic", *path);
574 warn = 1;
575 }
576 if (q[0] != '\r')
577 continue;
578 if (&q[1] == end) {
579 cr = 1;
580 break;
581 }
582 if (q[1] == '\n')
583 break;
584 }
585 if (write(fd, p, q - p) != q - p)
586 error("write('%s')", *path);
587 }
588 }
589 if (tty)
590 info(" \b\b");
591 if (text)
592 info(" (text)");
593 info("\n");
594
595 /* set access and modification time */
596 tv[0].tv_sec = now;
597 tv[0].tv_usec = 0;
598 tv[1].tv_sec = mtime;
599 tv[1].tv_usec = 0;
600 if (futimes(fd, tv) != 0)
601 error("utimes('%s')", *path);
602 if (close(fd) != 0)
603 error("close('%s')", *path);
604}
605
606/*
607 * Extract a zipfile entry: first perform some sanity checks to ensure
608 * that it is either a directory or a regular file and that the path is
609 * not absolute and does not try to break out of the current directory;
610 * then call either extract_dir() or extract_file() as appropriate.
611 *
612 * This is complicated a bit by the various ways in which we need to
613 * manipulate the path name. Case conversion (if requested by the -L
614 * option) happens first, but the include / exclude patterns are applied
615 * to the full converted path name, before the directory part of the path
616 * is removed in accordance with the -j option. Sanity checks are
617 * intentionally done earlier than they need to be, so the user will get a
618 * warning about insecure paths even for files or directories which
619 * wouldn't be extracted anyway.
620 */
621static void
622extract(struct archive *a, struct archive_entry *e)
623{
624 char *pathname, *realpathname;
625 mode_t filetype;
626 char *p, *q;
627
628 pathname = pathdup(archive_entry_pathname(e));
629 filetype = archive_entry_filetype(e);
630
631 /* sanity checks */
632 if (pathname[0] == '/' ||
633 strncmp(pathname, "../", 3) == 0 ||
634 strstr(pathname, "/../") != NULL) {
635 warningx("skipping insecure entry '%s'", pathname);
636 ac(archive_read_data_skip(a));
637 free(pathname);
638 return;
639 }
640
641 /* I don't think this can happen in a zipfile.. */
642 if (!S_ISDIR(filetype) && !S_ISREG(filetype)) {
643 warningx("skipping non-regular entry '%s'", pathname);
644 ac(archive_read_data_skip(a));
645 free(pathname);
646 return;
647 }
648
649 /* skip directories in -j case */
650 if (S_ISDIR(filetype) && j_opt) {
651 ac(archive_read_data_skip(a));
652 free(pathname);
653 return;
654 }
655
656 /* apply include / exclude patterns */
657 if (!accept_pathname(pathname)) {
658 ac(archive_read_data_skip(a));
659 free(pathname);
660 return;
661 }
662
663 /* apply -j and -d */
664 if (j_opt) {
665 for (p = q = pathname; *p; ++p)
666 if (*p == '/')
667 q = p + 1;
668 realpathname = pathcat(d_arg, q);
669 } else {
670 realpathname = pathcat(d_arg, pathname);
671 }
672
673 /* ensure that parent directory exists */
674 make_parent(realpathname);
675
676 if (S_ISDIR(filetype))
677 extract_dir(a, e, realpathname);
678 else
679 extract_file(a, e, &realpathname);
680
681 free(realpathname);
682 free(pathname);
683}
684
685static void
686extract_stdout(struct archive *a, struct archive_entry *e)
687{
688 char *pathname;
689 mode_t filetype;
690 int cr, text, warn;
691 ssize_t len;
692 unsigned char *p, *q, *end;
693
694 pathname = pathdup(archive_entry_pathname(e));
695 filetype = archive_entry_filetype(e);
696
697 /* I don't think this can happen in a zipfile.. */
698 if (!S_ISDIR(filetype) && !S_ISREG(filetype)) {
699 warningx("skipping non-regular entry '%s'", pathname);
700 ac(archive_read_data_skip(a));
701 free(pathname);
702 return;
703 }
704
705 /* skip directories in -j case */
706 if (S_ISDIR(filetype)) {
707 ac(archive_read_data_skip(a));
708 free(pathname);
709 return;
710 }
711
712 /* apply include / exclude patterns */
713 if (!accept_pathname(pathname)) {
714 ac(archive_read_data_skip(a));
715 free(pathname);
716 return;
717 }
718
719 if (c_opt)
720 info("x %s\n", pathname);
721
722 text = a_opt;
723 warn = 0;
724 cr = 0;
725 for (int n = 0; ; n++) {
726 len = archive_read_data(a, buffer, sizeof buffer);
727
728 if (len < 0)
729 ac(len);
730
731 /* left over CR from previous buffer */
732 if (a_opt && cr) {
733 if (len == 0 || buffer[0] != '\n') {
734 if (fwrite("\r", 1, 1, stderr) != 1)
735 error("write('%s')", pathname);
736 }
737 cr = 0;
738 }
739
740 /* EOF */
741 if (len == 0)
742 break;
743 end = buffer + len;
744
745 /*
746 * Detect whether this is a text file. The correct way to
747 * do this is to check the least significant bit of the
748 * "internal file attributes" field of the corresponding
749 * file header in the central directory, but libarchive
750 * does not read the central directory, so we have to
751 * guess by looking for non-ASCII characters in the
752 * buffer. Hopefully we won't guess wrong. If we do
753 * guess wrong, we print a warning message later.
754 */
755 if (a_opt && n == 0) {
756 for (p = buffer; p < end; ++p) {
757 if (!isascii((unsigned char)*p)) {
758 text = 0;
759 break;
760 }
761 }
762 }
763
764 /* simple case */
765 if (!a_opt || !text) {
766 if (fwrite(buffer, 1, len, stdout) != (size_t)len)
767 error("write('%s')", pathname);
768 continue;
769 }
770
771 /* hard case: convert \r\n to \n (sigh...) */
772 for (p = buffer; p < end; p = q + 1) {
773 for (q = p; q < end; q++) {
774 if (!warn && !isascii(*q)) {
775 warningx("%s may be corrupted due"
776 " to weak text file detection"
777 " heuristic", pathname);
778 warn = 1;
779 }
780 if (q[0] != '\r')
781 continue;
782 if (&q[1] == end) {
783 cr = 1;
784 break;
785 }
786 if (q[1] == '\n')
787 break;
788 }
789 if (fwrite(p, 1, q - p, stdout) != (size_t)(q - p))
790 error("write('%s')", pathname);
791 }
792 }
793
794 free(pathname);
795}
796
797/*
798 * Print the name of an entry to stdout.
799 */
800static void
801list(struct archive *a, struct archive_entry *e)
802{
803 char buf[20];
804 time_t mtime;
805
806 mtime = archive_entry_mtime(e);
807 strftime(buf, sizeof(buf), "%m-%d-%g %R", localtime(&mtime));
808
809 if (!zipinfo_mode) {
810 if (v_opt == 1) {
811 printf(" %8ju %s %s\n",
812 (uintmax_t)archive_entry_size(e),
813 buf, archive_entry_pathname(e));
814 } else if (v_opt == 2) {
815 printf("%8ju Stored %7ju 0%% %s %08x %s\n",
816 (uintmax_t)archive_entry_size(e),
817 (uintmax_t)archive_entry_size(e),
818 buf,
819 0U,
820 archive_entry_pathname(e));
821 }
822 } else {
823 if (Z1_opt)
824 printf("%s\n",archive_entry_pathname(e));
825 }
826 ac(archive_read_data_skip(a));
827}
828
829/*
830 * Extract to memory to check CRC
831 */
832static int
833test(struct archive *a, struct archive_entry *e)
834{
835 ssize_t len;
836 int error_count;
837
838 error_count = 0;
839 if (S_ISDIR(archive_entry_filetype(e)))
840 return 0;
841
842 info(" testing: %s\t", archive_entry_pathname(e));
843 while ((len = archive_read_data(a, buffer, sizeof buffer)) > 0)
844 /* nothing */;
845 if (len < 0) {
846 info(" %s\n", archive_error_string(a));
847 ++error_count;
848 } else {
849 info(" OK\n");
850 }
851
852 /* shouldn't be necessary, but it doesn't hurt */
853 ac(archive_read_data_skip(a));
854
855 return error_count;
856}
857
858
859/*
860 * Main loop: open the zipfile, iterate over its contents and decide what
861 * to do with each entry.
862 */
863static void
864unzip(const char *fn)
865{
866 struct archive *a;
867 struct archive_entry *e;
868 int ret;
869 uintmax_t total_size, file_count, error_count;
870
871 if ((a = archive_read_new()) == NULL)
872 error("archive_read_new failed");
873
874 ac(archive_read_support_format_zip(a));
875 ac(archive_read_open_filename(a, fn, 8192));
876
877 if (!zipinfo_mode) {
878 if (!p_opt && !q_opt)
879 printf("Archive: %s\n", fn);
880 if (v_opt == 1) {
881 printf(" Length Date Time Name\n");
882 printf(" -------- ---- ---- ----\n");
883 } else if (v_opt == 2) {
884 printf(" Length Method Size Ratio Date Time CRC-32 Name\n");
885 printf("-------- ------ ------- ----- ---- ---- ------ ----\n");
886 }
887 }
888
889 total_size = 0;
890 file_count = 0;
891 error_count = 0;
892 for (;;) {
893 ret = archive_read_next_header(a, &e);
894 if (ret == ARCHIVE_EOF)
895 break;
896 ac(ret);
897 if (!zipinfo_mode) {
898 if (t_opt)
899 error_count += test(a, e);
900 else if (v_opt)
901 list(a, e);
902 else if (p_opt || c_opt)
903 extract_stdout(a, e);
904 else
905 extract(a, e);
906 } else {
907 if (Z1_opt)
908 list(a, e);
909 }
910
911 total_size += archive_entry_size(e);
912 ++file_count;
913 }
914
915 if (zipinfo_mode) {
916 if (v_opt == 1) {
917 printf(" -------- -------\n");
918 printf(" %8ju %ju file%s\n",
919 total_size, file_count, file_count != 1 ? "s" : "");
920 } else if (v_opt == 2) {
921 printf("-------- ------- --- -------\n");
922 printf("%8ju %7ju 0%% %ju file%s\n",
923 total_size, total_size, file_count,
924 file_count != 1 ? "s" : "");
925 }
926 }
927
928 ac(archive_read_close(a));
929 (void)archive_read_finish(a);
929 (void)archive_read_free(a);
930
931 if (t_opt) {
932 if (error_count > 0) {
933 errorx("%d checksum error(s) found.", error_count);
934 }
935 else {
936 printf("No errors detected in compressed data of %s.\n",
937 fn);
938 }
939 }
940}
941
942static void
943usage(void)
944{
945
946 fprintf(stderr, "usage: unzip [-aCcfjLlnopqtuvZ1] [-d dir] [-x pattern] zipfile\n");
947 exit(1);
948}
949
950static int
951getopts(int argc, char *argv[])
952{
953 int opt;
954
955 optreset = optind = 1;
956 while ((opt = getopt(argc, argv, "aCcd:fjLlnopqtuvx:Z1")) != -1)
957 switch (opt) {
958 case '1':
959 Z1_opt = 1;
960 break;
961 case 'a':
962 a_opt = 1;
963 break;
964 case 'C':
965 C_opt = 1;
966 break;
967 case 'c':
968 c_opt = 1;
969 break;
970 case 'd':
971 d_arg = optarg;
972 break;
973 case 'f':
974 f_opt = 1;
975 break;
976 case 'j':
977 j_opt = 1;
978 break;
979 case 'L':
980 L_opt = 1;
981 break;
982 case 'l':
983 if (v_opt == 0)
984 v_opt = 1;
985 break;
986 case 'n':
987 n_opt = 1;
988 break;
989 case 'o':
990 o_opt = 1;
991 q_opt = 1;
992 break;
993 case 'p':
994 p_opt = 1;
995 break;
996 case 'q':
997 q_opt = 1;
998 break;
999 case 't':
1000 t_opt = 1;
1001 break;
1002 case 'u':
1003 u_opt = 1;
1004 break;
1005 case 'v':
1006 v_opt = 2;
1007 break;
1008 case 'x':
1009 add_pattern(&exclude, optarg);
1010 break;
1011 case 'Z':
1012 zipinfo_mode = 1;
1013 break;
1014 default:
1015 usage();
1016 }
1017
1018 return (optind);
1019}
1020
1021int
1022main(int argc, char *argv[])
1023{
1024 const char *zipfile;
1025 int nopts;
1026
1027 if (isatty(STDOUT_FILENO))
1028 tty = 1;
1029
1030 if (getenv("UNZIP_DEBUG") != NULL)
1031 unzip_debug = 1;
1032 for (int i = 0; i < argc; ++i)
1033 debug("%s%c", argv[i], (i < argc - 1) ? ' ' : '\n');
1034
1035 /*
1036 * Info-ZIP's unzip(1) expects certain options to come before the
1037 * zipfile name, and others to come after - though it does not
1038 * enforce this. For simplicity, we accept *all* options both
1039 * before and after the zipfile name.
1040 */
1041 nopts = getopts(argc, argv);
1042
1043 /*
1044 * When more of the zipinfo mode options are implemented, this
1045 * will need to change.
1046 */
1047 if (zipinfo_mode && !Z1_opt) {
1048 printf("Zipinfo mode needs additional options\n");
1049 exit(1);
1050 }
1051
1052 if (argc <= nopts)
1053 usage();
1054 zipfile = argv[nopts++];
1055
1056 if (strcmp(zipfile, "-") == 0)
1057 zipfile = NULL; /* STDIN */
1058
1059 while (nopts < argc && *argv[nopts] != '-')
1060 add_pattern(&include, argv[nopts++]);
1061
1062 nopts--; /* fake argv[0] */
1063 nopts += getopts(argc - nopts, argv + nopts);
1064
1065 if (n_opt + o_opt + u_opt > 1)
1066 errorx("-n, -o and -u are contradictory");
1067
1068 time(&now);
1069
1070 unzip(zipfile);
1071
1072 exit(0);
1073}
930
931 if (t_opt) {
932 if (error_count > 0) {
933 errorx("%d checksum error(s) found.", error_count);
934 }
935 else {
936 printf("No errors detected in compressed data of %s.\n",
937 fn);
938 }
939 }
940}
941
942static void
943usage(void)
944{
945
946 fprintf(stderr, "usage: unzip [-aCcfjLlnopqtuvZ1] [-d dir] [-x pattern] zipfile\n");
947 exit(1);
948}
949
950static int
951getopts(int argc, char *argv[])
952{
953 int opt;
954
955 optreset = optind = 1;
956 while ((opt = getopt(argc, argv, "aCcd:fjLlnopqtuvx:Z1")) != -1)
957 switch (opt) {
958 case '1':
959 Z1_opt = 1;
960 break;
961 case 'a':
962 a_opt = 1;
963 break;
964 case 'C':
965 C_opt = 1;
966 break;
967 case 'c':
968 c_opt = 1;
969 break;
970 case 'd':
971 d_arg = optarg;
972 break;
973 case 'f':
974 f_opt = 1;
975 break;
976 case 'j':
977 j_opt = 1;
978 break;
979 case 'L':
980 L_opt = 1;
981 break;
982 case 'l':
983 if (v_opt == 0)
984 v_opt = 1;
985 break;
986 case 'n':
987 n_opt = 1;
988 break;
989 case 'o':
990 o_opt = 1;
991 q_opt = 1;
992 break;
993 case 'p':
994 p_opt = 1;
995 break;
996 case 'q':
997 q_opt = 1;
998 break;
999 case 't':
1000 t_opt = 1;
1001 break;
1002 case 'u':
1003 u_opt = 1;
1004 break;
1005 case 'v':
1006 v_opt = 2;
1007 break;
1008 case 'x':
1009 add_pattern(&exclude, optarg);
1010 break;
1011 case 'Z':
1012 zipinfo_mode = 1;
1013 break;
1014 default:
1015 usage();
1016 }
1017
1018 return (optind);
1019}
1020
1021int
1022main(int argc, char *argv[])
1023{
1024 const char *zipfile;
1025 int nopts;
1026
1027 if (isatty(STDOUT_FILENO))
1028 tty = 1;
1029
1030 if (getenv("UNZIP_DEBUG") != NULL)
1031 unzip_debug = 1;
1032 for (int i = 0; i < argc; ++i)
1033 debug("%s%c", argv[i], (i < argc - 1) ? ' ' : '\n');
1034
1035 /*
1036 * Info-ZIP's unzip(1) expects certain options to come before the
1037 * zipfile name, and others to come after - though it does not
1038 * enforce this. For simplicity, we accept *all* options both
1039 * before and after the zipfile name.
1040 */
1041 nopts = getopts(argc, argv);
1042
1043 /*
1044 * When more of the zipinfo mode options are implemented, this
1045 * will need to change.
1046 */
1047 if (zipinfo_mode && !Z1_opt) {
1048 printf("Zipinfo mode needs additional options\n");
1049 exit(1);
1050 }
1051
1052 if (argc <= nopts)
1053 usage();
1054 zipfile = argv[nopts++];
1055
1056 if (strcmp(zipfile, "-") == 0)
1057 zipfile = NULL; /* STDIN */
1058
1059 while (nopts < argc && *argv[nopts] != '-')
1060 add_pattern(&include, argv[nopts++]);
1061
1062 nopts--; /* fake argv[0] */
1063 nopts += getopts(argc - nopts, argv + nopts);
1064
1065 if (n_opt + o_opt + u_opt > 1)
1066 errorx("-n, -o and -u are contradictory");
1067
1068 time(&now);
1069
1070 unzip(zipfile);
1071
1072 exit(0);
1073}