xref: /freebsd-13.1/usr.bin/unzip/unzip.c (revision 3bfe2131)
1 /*-
2  * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
3  *
4  * Copyright (c) 2009, 2010 Joerg Sonnenberger <[email protected]>
5  * Copyright (c) 2007-2008 Dag-Erling Smørgrav
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  * $FreeBSD$
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/queue.h>
40 #include <sys/stat.h>
41 
42 #include <ctype.h>
43 #include <errno.h>
44 #include <fcntl.h>
45 #include <fnmatch.h>
46 #include <stdarg.h>
47 #include <stdio.h>
48 #include <stdlib.h>
49 #include <string.h>
50 #include <unistd.h>
51 
52 #include <archive.h>
53 #include <archive_entry.h>
54 #include <readpassphrase.h>
55 
56 /* command-line options */
57 static int		 a_opt;		/* convert EOL */
58 static int		 C_opt;		/* match case-insensitively */
59 static int		 c_opt;		/* extract to stdout */
60 static const char	*d_arg;		/* directory */
61 static int		 f_opt;		/* update existing files only */
62 static int		 j_opt;		/* junk directories */
63 static int		 L_opt;		/* lowercase names */
64 static int		 n_opt;		/* never overwrite */
65 static int		 o_opt;		/* always overwrite */
66 static int		 p_opt;		/* extract to stdout, quiet */
67 static char		*P_arg;		/* passphrase */
68 static int		 q_opt;		/* quiet */
69 static int		 t_opt;		/* test */
70 static int		 u_opt;		/* update */
71 static int		 v_opt;		/* verbose/list */
72 static const char	*y_str = "";	/* 4 digit year */
73 static int		 Z1_opt;	/* zipinfo mode list files only */
74 
75 /* debug flag */
76 static int		 unzip_debug;
77 
78 /* zipinfo mode */
79 static int		 zipinfo_mode;
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 /* for an interactive passphrase input */
101 static char *passphrase_buf;
102 
103 /* fatal error message + errno */
104 static void
error(const char * fmt,...)105 error(const char *fmt, ...)
106 {
107 	va_list ap;
108 
109 	if (noeol)
110 		fprintf(stdout, "\n");
111 	fflush(stdout);
112 	fprintf(stderr, "unzip: ");
113 	va_start(ap, fmt);
114 	vfprintf(stderr, fmt, ap);
115 	va_end(ap);
116 	fprintf(stderr, ": %s\n", strerror(errno));
117 	exit(EXIT_FAILURE);
118 }
119 
120 /* fatal error message, no errno */
121 static void
errorx(const char * fmt,...)122 errorx(const char *fmt, ...)
123 {
124 	va_list ap;
125 
126 	if (noeol)
127 		fprintf(stdout, "\n");
128 	fflush(stdout);
129 	fprintf(stderr, "unzip: ");
130 	va_start(ap, fmt);
131 	vfprintf(stderr, fmt, ap);
132 	va_end(ap);
133 	fprintf(stderr, "\n");
134 	exit(EXIT_FAILURE);
135 }
136 
137 /* non-fatal error message + errno */
138 static void
warning(const char * fmt,...)139 warning(const char *fmt, ...)
140 {
141 	va_list ap;
142 
143 	if (noeol)
144 		fprintf(stdout, "\n");
145 	fflush(stdout);
146 	fprintf(stderr, "unzip: ");
147 	va_start(ap, fmt);
148 	vfprintf(stderr, fmt, ap);
149 	va_end(ap);
150 	fprintf(stderr, ": %s\n", strerror(errno));
151 }
152 
153 /* non-fatal error message, no errno */
154 static void
warningx(const char * fmt,...)155 warningx(const char *fmt, ...)
156 {
157 	va_list ap;
158 
159 	if (noeol)
160 		fprintf(stdout, "\n");
161 	fflush(stdout);
162 	fprintf(stderr, "unzip: ");
163 	va_start(ap, fmt);
164 	vfprintf(stderr, fmt, ap);
165 	va_end(ap);
166 	fprintf(stderr, "\n");
167 }
168 
169 /* informational message (if not -q) */
170 static void
info(const char * fmt,...)171 info(const char *fmt, ...)
172 {
173 	va_list ap;
174 
175 	if (q_opt && !unzip_debug)
176 		return;
177 	va_start(ap, fmt);
178 	vfprintf(stdout, fmt, ap);
179 	va_end(ap);
180 	fflush(stdout);
181 
182 	if (*fmt == '\0')
183 		noeol = 1;
184 	else
185 		noeol = fmt[strlen(fmt) - 1] != '\n';
186 }
187 
188 /* debug message (if unzip_debug) */
189 static void
debug(const char * fmt,...)190 debug(const char *fmt, ...)
191 {
192 	va_list ap;
193 
194 	if (!unzip_debug)
195 		return;
196 	va_start(ap, fmt);
197 	vfprintf(stderr, fmt, ap);
198 	va_end(ap);
199 	fflush(stderr);
200 
201 	if (*fmt == '\0')
202 		noeol = 1;
203 	else
204 		noeol = fmt[strlen(fmt) - 1] != '\n';
205 }
206 
207 /* duplicate a path name, possibly converting to lower case */
208 static char *
pathdup(const char * path)209 pathdup(const char *path)
210 {
211 	char *str;
212 	size_t i, len;
213 
214 	if (path == NULL || path[0] == '\0')
215 		return (NULL);
216 
217 	len = strlen(path);
218 	while (len && path[len - 1] == '/')
219 		len--;
220 	if ((str = malloc(len + 1)) == NULL) {
221 		errno = ENOMEM;
222 		error("malloc()");
223 	}
224 	if (L_opt) {
225 		for (i = 0; i < len; ++i)
226 			str[i] = tolower((unsigned char)path[i]);
227 	} else {
228 		memcpy(str, path, len);
229 	}
230 	str[len] = '\0';
231 
232 	return (str);
233 }
234 
235 /* concatenate two path names */
236 static char *
pathcat(const char * prefix,const char * path)237 pathcat(const char *prefix, const char *path)
238 {
239 	char *str;
240 	size_t prelen, len;
241 
242 	prelen = prefix ? strlen(prefix) + 1 : 0;
243 	len = strlen(path) + 1;
244 	if ((str = malloc(prelen + len)) == NULL) {
245 		errno = ENOMEM;
246 		error("malloc()");
247 	}
248 	if (prefix) {
249 		memcpy(str, prefix, prelen);	/* includes zero */
250 		str[prelen - 1] = '/';		/* splat zero */
251 	}
252 	memcpy(str + prelen, path, len);	/* includes zero */
253 
254 	return (str);
255 }
256 
257 /*
258  * Pattern lists for include / exclude processing
259  */
260 struct pattern {
261 	STAILQ_ENTRY(pattern) link;
262 	char pattern[];
263 };
264 
265 STAILQ_HEAD(pattern_list, pattern);
266 static struct pattern_list include = STAILQ_HEAD_INITIALIZER(include);
267 static struct pattern_list exclude = STAILQ_HEAD_INITIALIZER(exclude);
268 
269 /*
270  * Add an entry to a pattern list
271  */
272 static void
add_pattern(struct pattern_list * list,const char * pattern)273 add_pattern(struct pattern_list *list, const char *pattern)
274 {
275 	struct pattern *entry;
276 	size_t len;
277 
278 	debug("adding pattern '%s'\n", pattern);
279 	len = strlen(pattern);
280 	if ((entry = malloc(sizeof *entry + len + 1)) == NULL) {
281 		errno = ENOMEM;
282 		error("malloc()");
283 	}
284 	memcpy(entry->pattern, pattern, len + 1);
285 	STAILQ_INSERT_TAIL(list, entry, link);
286 }
287 
288 /*
289  * Match a string against a list of patterns
290  */
291 static int
match_pattern(struct pattern_list * list,const char * str)292 match_pattern(struct pattern_list *list, const char *str)
293 {
294 	struct pattern *entry;
295 
296 	STAILQ_FOREACH(entry, list, link) {
297 		if (fnmatch(entry->pattern, str, C_opt ? FNM_CASEFOLD : 0) == 0)
298 			return (1);
299 	}
300 	return (0);
301 }
302 
303 /*
304  * Verify that a given pathname is in the include list and not in the
305  * exclude list.
306  */
307 static int
accept_pathname(const char * pathname)308 accept_pathname(const char *pathname)
309 {
310 
311 	if (!STAILQ_EMPTY(&include) && !match_pattern(&include, pathname))
312 		return (0);
313 	if (!STAILQ_EMPTY(&exclude) && match_pattern(&exclude, pathname))
314 		return (0);
315 	return (1);
316 }
317 
318 /*
319  * Create the specified directory with the specified mode, taking certain
320  * precautions on they way.
321  */
322 static void
make_dir(const char * path,int mode)323 make_dir(const char *path, int mode)
324 {
325 	struct stat sb;
326 
327 	if (lstat(path, &sb) == 0) {
328 		if (S_ISDIR(sb.st_mode))
329 			return;
330 		/*
331 		 * Normally, we should either ask the user about removing
332 		 * the non-directory of the same name as a directory we
333 		 * wish to create, or respect the -n or -o command-line
334 		 * options.  However, this may lead to a later failure or
335 		 * even compromise (if this non-directory happens to be a
336 		 * symlink to somewhere unsafe), so we don't.
337 		 */
338 
339 		/*
340 		 * Don't check unlink() result; failure will cause mkdir()
341 		 * to fail later, which we will catch.
342 		 */
343 		(void)unlink(path);
344 	}
345 	if (mkdir(path, mode) != 0 && errno != EEXIST)
346 		error("mkdir('%s')", path);
347 }
348 
349 /*
350  * Ensure that all directories leading up to (but not including) the
351  * specified path exist.
352  *
353  * XXX inefficient + modifies the file in-place
354  */
355 static void
make_parent(char * path)356 make_parent(char *path)
357 {
358 	struct stat sb;
359 	char *sep;
360 
361 	sep = strrchr(path, '/');
362 	if (sep == NULL || sep == path)
363 		return;
364 	*sep = '\0';
365 	if (lstat(path, &sb) == 0) {
366 		if (S_ISDIR(sb.st_mode)) {
367 			*sep = '/';
368 			return;
369 		}
370 		unlink(path);
371 	}
372 	make_parent(path);
373 	mkdir(path, 0755);
374 	*sep = '/';
375 
376 #if 0
377 	for (sep = path; (sep = strchr(sep, '/')) != NULL; sep++) {
378 		/* root in case of absolute d_arg */
379 		if (sep == path)
380 			continue;
381 		*sep = '\0';
382 		make_dir(path, 0755);
383 		*sep = '/';
384 	}
385 #endif
386 }
387 
388 /*
389  * Extract a directory.
390  */
391 static void
extract_dir(struct archive * a,struct archive_entry * e,const char * path)392 extract_dir(struct archive *a, struct archive_entry *e, const char *path)
393 {
394 	int mode;
395 
396 	/*
397 	 * Dropbox likes to create '/' directory entries, just ignore
398 	 * such junk.
399 	 */
400 	if (*path == '\0')
401 		return;
402 
403 	mode = archive_entry_mode(e) & 0777;
404 	if (mode == 0)
405 		mode = 0755;
406 
407 	/*
408 	 * Some zipfiles contain directories with weird permissions such
409 	 * as 0644 or 0444.  This can cause strange issues such as being
410 	 * unable to extract files into the directory we just created, or
411 	 * the user being unable to remove the directory later without
412 	 * first manually changing its permissions.  Therefore, we whack
413 	 * the permissions into shape, assuming that the user wants full
414 	 * access and that anyone who gets read access also gets execute
415 	 * access.
416 	 */
417 	mode |= 0700;
418 	if (mode & 0040)
419 		mode |= 0010;
420 	if (mode & 0004)
421 		mode |= 0001;
422 
423 	info("   creating: %s/\n", path);
424 	make_dir(path, mode);
425 	ac(archive_read_data_skip(a));
426 }
427 
428 static unsigned char buffer[8192];
429 static char spinner[] = { '|', '/', '-', '\\' };
430 
431 static int
handle_existing_file(char ** path)432 handle_existing_file(char **path)
433 {
434 	size_t alen;
435 	ssize_t len;
436 	char buf[4];
437 
438 	for (;;) {
439 		fprintf(stderr,
440 		    "replace %s? [y]es, [n]o, [A]ll, [N]one, [r]ename: ",
441 		    *path);
442 		if (fgets(buf, sizeof(buf), stdin) == NULL) {
443 			clearerr(stdin);
444 			printf("NULL\n(EOF or read error, "
445 			    "treating as \"[N]one\"...)\n");
446 			n_opt = 1;
447 			return -1;
448 		}
449 		switch (*buf) {
450 		case 'A':
451 			o_opt = 1;
452 			/* FALLTHROUGH */
453 		case 'y':
454 		case 'Y':
455 			(void)unlink(*path);
456 			return 1;
457 		case 'N':
458 			n_opt = 1;
459 			/* FALLTHROUGH */
460 		case 'n':
461 			return -1;
462 		case 'r':
463 		case 'R':
464 			printf("New name: ");
465 			fflush(stdout);
466 			free(*path);
467 			*path = NULL;
468 			alen = 0;
469 			len = getline(path, &alen, stdin);
470 			if ((*path)[len - 1] == '\n')
471 				(*path)[len - 1] = '\0';
472 			return 0;
473 		default:
474 			break;
475 		}
476 	}
477 }
478 
479 /*
480  * Detect binary files by a combination of character white list and
481  * black list. NUL bytes and other control codes without use in text files
482  * result directly in switching the file to binary mode. Otherwise, at least
483  * one white-listed byte has to be found.
484  *
485  * Black-listed: 0..6, 14..25, 28..31
486  * 0xf3ffc07f = 11110011111111111100000001111111b
487  * White-listed: 9..10, 13, >= 32
488  * 0x00002600 = 00000000000000000010011000000000b
489  *
490  * See the proginfo/txtvsbin.txt in the zip sources for a detailed discussion.
491  */
492 #define BYTE_IS_BINARY(x)	((x) < 32 && (0xf3ffc07fU & (1U << (x))))
493 #define	BYTE_IS_TEXT(x)		((x) >= 32 || (0x00002600U & (1U << (x))))
494 
495 static int
check_binary(const unsigned char * buf,size_t len)496 check_binary(const unsigned char *buf, size_t len)
497 {
498 	int rv;
499 	for (rv = 1; len--; ++buf) {
500 		if (BYTE_IS_BINARY(*buf))
501 			return 1;
502 		if (BYTE_IS_TEXT(*buf))
503 			rv = 0;
504 	}
505 
506 	return rv;
507 }
508 
509 /*
510  * Extract to a file descriptor
511  */
512 static int
extract2fd(struct archive * a,char * pathname,int fd)513 extract2fd(struct archive *a, char *pathname, int fd)
514 {
515 	int cr, text, warn;
516 	ssize_t len;
517 	unsigned char *p, *q, *end;
518 
519 	text = a_opt;
520 	warn = 0;
521 	cr = 0;
522 
523 	/* loop over file contents and write to fd */
524 	for (int n = 0; ; n++) {
525 		if (fd != STDOUT_FILENO)
526 			if (tty && (n % 4) == 0)
527 				info(" %c\b\b", spinner[(n / 4) % sizeof spinner]);
528 
529 		len = archive_read_data(a, buffer, sizeof buffer);
530 
531 		if (len < 0)
532 			ac(len);
533 
534 		/* left over CR from previous buffer */
535 		if (a_opt && cr) {
536 			if (len == 0 || buffer[0] != '\n')
537 				if (write(fd, "\r", 1) != 1)
538 					error("write('%s')", pathname);
539 			cr = 0;
540 		}
541 
542 		/* EOF */
543 		if (len == 0)
544 			break;
545 		end = buffer + len;
546 
547 		/*
548 		 * Detect whether this is a text file.  The correct way to
549 		 * do this is to check the least significant bit of the
550 		 * "internal file attributes" field of the corresponding
551 		 * file header in the central directory, but libarchive
552 		 * does not provide access to this field, so we have to
553 		 * guess by looking for non-ASCII characters in the
554 		 * buffer.  Hopefully we won't guess wrong.  If we do
555 		 * guess wrong, we print a warning message later.
556 		 */
557 		if (a_opt && n == 0) {
558 			if (check_binary(buffer, len))
559 				text = 0;
560 		}
561 
562 		/* simple case */
563 		if (!a_opt || !text) {
564 			if (write(fd, buffer, len) != len)
565 				error("write('%s')", pathname);
566 			continue;
567 		}
568 
569 		/* hard case: convert \r\n to \n (sigh...) */
570 		for (p = buffer; p < end; p = q + 1) {
571 			for (q = p; q < end; q++) {
572 				if (!warn && BYTE_IS_BINARY(*q)) {
573 					warningx("%s may be corrupted due"
574 					    " to weak text file detection"
575 					    " heuristic", pathname);
576 					warn = 1;
577 				}
578 				if (q[0] != '\r')
579 					continue;
580 				if (&q[1] == end) {
581 					cr = 1;
582 					break;
583 				}
584 				if (q[1] == '\n')
585 					break;
586 			}
587 			if (write(fd, p, q - p) != q - p)
588 				error("write('%s')", pathname);
589 		}
590 	}
591 
592 	return text;
593 }
594 
595 /*
596  * Extract a regular file.
597  */
598 static void
extract_file(struct archive * a,struct archive_entry * e,char ** path)599 extract_file(struct archive *a, struct archive_entry *e, char **path)
600 {
601 	int mode;
602 	struct timespec mtime;
603 	struct stat sb;
604 	struct timespec ts[2];
605 	int fd, check, text;
606 	const char *linkname;
607 
608 	mode = archive_entry_mode(e) & 0777;
609 	if (mode == 0)
610 		mode = 0644;
611 	mtime.tv_sec = archive_entry_mtime(e);
612 	mtime.tv_nsec = archive_entry_mtime_nsec(e);
613 
614 	/* look for existing file of same name */
615 recheck:
616 	if (lstat(*path, &sb) == 0) {
617 		if (u_opt || f_opt) {
618 			/* check if up-to-date */
619 			if (S_ISREG(sb.st_mode) &&
620 			    (sb.st_mtim.tv_sec > mtime.tv_sec ||
621 			    (sb.st_mtim.tv_sec == mtime.tv_sec &&
622 			    sb.st_mtim.tv_nsec >= mtime.tv_nsec)))
623 				return;
624 			(void)unlink(*path);
625 		} else if (o_opt) {
626 			/* overwrite */
627 			(void)unlink(*path);
628 		} else if (n_opt) {
629 			/* do not overwrite */
630 			return;
631 		} else {
632 			check = handle_existing_file(path);
633 			if (check == 0)
634 				goto recheck;
635 			if (check == -1)
636 				return; /* do not overwrite */
637 		}
638 	} else {
639 		if (f_opt)
640 			return;
641 	}
642 
643 	ts[0].tv_sec = 0;
644 	ts[0].tv_nsec = UTIME_NOW;
645 	ts[1] = mtime;
646 
647 	/* process symlinks */
648 	linkname = archive_entry_symlink(e);
649 	if (linkname != NULL) {
650 		if (symlink(linkname, *path) != 0)
651 			error("symlink('%s')", *path);
652 		info(" extracting: %s -> %s\n", *path, linkname);
653 		if (lchmod(*path, mode) != 0)
654 			warning("Cannot set mode for '%s'", *path);
655 		/* set access and modification time */
656 		if (utimensat(AT_FDCWD, *path, ts, AT_SYMLINK_NOFOLLOW) != 0)
657 			warning("utimensat('%s')", *path);
658 		return;
659 	}
660 
661 	if ((fd = open(*path, O_RDWR|O_CREAT|O_TRUNC, mode)) < 0)
662 		error("open('%s')", *path);
663 
664 	info(" extracting: %s", *path);
665 
666 	text = extract2fd(a, *path, fd);
667 
668 	if (tty)
669 		info("  \b\b");
670 	if (text)
671 		info(" (text)");
672 	info("\n");
673 
674 	/* set access and modification time */
675 	if (futimens(fd, ts) != 0)
676 		error("futimens('%s')", *path);
677 	if (close(fd) != 0)
678 		error("close('%s')", *path);
679 }
680 
681 /*
682  * Extract a zipfile entry: first perform some sanity checks to ensure
683  * that it is either a directory or a regular file and that the path is
684  * not absolute and does not try to break out of the current directory;
685  * then call either extract_dir() or extract_file() as appropriate.
686  *
687  * This is complicated a bit by the various ways in which we need to
688  * manipulate the path name.  Case conversion (if requested by the -L
689  * option) happens first, but the include / exclude patterns are applied
690  * to the full converted path name, before the directory part of the path
691  * is removed in accordance with the -j option.  Sanity checks are
692  * intentionally done earlier than they need to be, so the user will get a
693  * warning about insecure paths even for files or directories which
694  * wouldn't be extracted anyway.
695  */
696 static void
extract(struct archive * a,struct archive_entry * e)697 extract(struct archive *a, struct archive_entry *e)
698 {
699 	char *pathname, *realpathname;
700 	mode_t filetype;
701 	char *p, *q;
702 
703 	if ((pathname = pathdup(archive_entry_pathname(e))) == NULL) {
704 		warningx("skipping empty or unreadable filename entry");
705 		ac(archive_read_data_skip(a));
706 		return;
707 	}
708 	filetype = archive_entry_filetype(e);
709 
710 	/* sanity checks */
711 	if (pathname[0] == '/' ||
712 	    strncmp(pathname, "../", 3) == 0 ||
713 	    strstr(pathname, "/../") != NULL) {
714 		warningx("skipping insecure entry '%s'", pathname);
715 		ac(archive_read_data_skip(a));
716 		free(pathname);
717 		return;
718 	}
719 
720 	/* I don't think this can happen in a zipfile.. */
721 	if (!S_ISDIR(filetype) && !S_ISREG(filetype) && !S_ISLNK(filetype)) {
722 		warningx("skipping non-regular entry '%s'", pathname);
723 		ac(archive_read_data_skip(a));
724 		free(pathname);
725 		return;
726 	}
727 
728 	/* skip directories in -j case */
729 	if (S_ISDIR(filetype) && j_opt) {
730 		ac(archive_read_data_skip(a));
731 		free(pathname);
732 		return;
733 	}
734 
735 	/* apply include / exclude patterns */
736 	if (!accept_pathname(pathname)) {
737 		ac(archive_read_data_skip(a));
738 		free(pathname);
739 		return;
740 	}
741 
742 	/* apply -j and -d */
743 	if (j_opt) {
744 		for (p = q = pathname; *p; ++p)
745 			if (*p == '/')
746 				q = p + 1;
747 		realpathname = pathcat(d_arg, q);
748 	} else {
749 		realpathname = pathcat(d_arg, pathname);
750 	}
751 
752 	/* ensure that parent directory exists */
753 	make_parent(realpathname);
754 
755 	if (S_ISDIR(filetype))
756 		extract_dir(a, e, realpathname);
757 	else
758 		extract_file(a, e, &realpathname);
759 
760 	free(realpathname);
761 	free(pathname);
762 }
763 
764 static void
extract_stdout(struct archive * a,struct archive_entry * e)765 extract_stdout(struct archive *a, struct archive_entry *e)
766 {
767 	char *pathname;
768 	mode_t filetype;
769 
770 	if ((pathname = pathdup(archive_entry_pathname(e))) == NULL) {
771 		warningx("skipping empty or unreadable filename entry");
772 		ac(archive_read_data_skip(a));
773 		return;
774 	}
775 	filetype = archive_entry_filetype(e);
776 
777 	/* I don't think this can happen in a zipfile.. */
778 	if (!S_ISDIR(filetype) && !S_ISREG(filetype) && !S_ISLNK(filetype)) {
779 		warningx("skipping non-regular entry '%s'", pathname);
780 		ac(archive_read_data_skip(a));
781 		free(pathname);
782 		return;
783 	}
784 
785 	/* skip directories in -j case */
786 	if (S_ISDIR(filetype)) {
787 		ac(archive_read_data_skip(a));
788 		free(pathname);
789 		return;
790 	}
791 
792 	/* apply include / exclude patterns */
793 	if (!accept_pathname(pathname)) {
794 		ac(archive_read_data_skip(a));
795 		free(pathname);
796 		return;
797 	}
798 
799 	if (c_opt)
800 		info("x %s\n", pathname);
801 
802 	(void)extract2fd(a, pathname, STDOUT_FILENO);
803 
804 	free(pathname);
805 }
806 
807 /*
808  * Print the name of an entry to stdout.
809  */
810 static void
list(struct archive * a,struct archive_entry * e)811 list(struct archive *a, struct archive_entry *e)
812 {
813 	char buf[20];
814 	time_t mtime;
815 	struct tm *tm;
816 
817 	mtime = archive_entry_mtime(e);
818 	tm = localtime(&mtime);
819 	if (*y_str)
820 		strftime(buf, sizeof(buf), "%m-%d-%G %R", tm);
821 	else
822 		strftime(buf, sizeof(buf), "%m-%d-%g %R", tm);
823 
824 	if (!zipinfo_mode) {
825 		if (v_opt == 1) {
826 			printf(" %8ju  %s   %s\n",
827 			    (uintmax_t)archive_entry_size(e),
828 			    buf, archive_entry_pathname(e));
829 		} else if (v_opt == 2) {
830 			printf("%8ju  Stored  %7ju   0%%  %s  %08x  %s\n",
831 			    (uintmax_t)archive_entry_size(e),
832 			    (uintmax_t)archive_entry_size(e),
833 			    buf,
834 			    0U,
835 			    archive_entry_pathname(e));
836 		}
837 	} else {
838 		if (Z1_opt)
839 			printf("%s\n",archive_entry_pathname(e));
840 	}
841 	ac(archive_read_data_skip(a));
842 }
843 
844 /*
845  * Extract to memory to check CRC
846  */
847 static int
test(struct archive * a,struct archive_entry * e)848 test(struct archive *a, struct archive_entry *e)
849 {
850 	ssize_t len;
851 	int error_count;
852 
853 	error_count = 0;
854 	if (S_ISDIR(archive_entry_filetype(e)))
855 		return 0;
856 
857 	info("    testing: %s\t", archive_entry_pathname(e));
858 	while ((len = archive_read_data(a, buffer, sizeof buffer)) > 0)
859 		/* nothing */;
860 	if (len < 0) {
861 		info(" %s\n", archive_error_string(a));
862 		++error_count;
863 	} else {
864 		info(" OK\n");
865 	}
866 
867 	/* shouldn't be necessary, but it doesn't hurt */
868 	ac(archive_read_data_skip(a));
869 
870 	return error_count;
871 }
872 
873 /*
874  * Callback function for reading passphrase.
875  * Originally from cpio.c and passphrase.c, libarchive.
876  */
877 #define PPBUFF_SIZE 1024
878 static const char *
passphrase_callback(struct archive * a,void * _client_data)879 passphrase_callback(struct archive *a, void *_client_data)
880 {
881 	char *p;
882 
883 	(void)a; /* UNUSED */
884 	(void)_client_data; /* UNUSED */
885 
886 	if (passphrase_buf == NULL) {
887 		passphrase_buf = malloc(PPBUFF_SIZE);
888 		if (passphrase_buf == NULL) {
889 			errno = ENOMEM;
890 			error("malloc()");
891 		}
892 	}
893 
894 	p = readpassphrase("\nEnter password: ", passphrase_buf,
895 		PPBUFF_SIZE, RPP_ECHO_OFF);
896 
897 	if (p == NULL && errno != EINTR)
898 		error("Error reading password");
899 
900 	return p;
901 }
902 
903 /*
904  * Main loop: open the zipfile, iterate over its contents and decide what
905  * to do with each entry.
906  */
907 static void
unzip(const char * fn)908 unzip(const char *fn)
909 {
910 	struct archive *a;
911 	struct archive_entry *e;
912 	int ret;
913 	uintmax_t total_size, file_count, error_count;
914 
915 	if ((a = archive_read_new()) == NULL)
916 		error("archive_read_new failed");
917 
918 	ac(archive_read_support_format_zip(a));
919 
920 	if (P_arg)
921 		archive_read_add_passphrase(a, P_arg);
922 	else
923 		archive_read_set_passphrase_callback(a, NULL,
924 			&passphrase_callback);
925 
926 	ac(archive_read_open_filename(a, fn, 8192));
927 
928 	if (!zipinfo_mode) {
929 		if (!p_opt && !q_opt)
930 			printf("Archive:  %s\n", fn);
931 		if (v_opt == 1) {
932 			printf("  Length     %sDate   Time    Name\n", y_str);
933 			printf(" --------    %s----   ----    ----\n", y_str);
934 		} else if (v_opt == 2) {
935 			printf(" Length   Method    Size  Ratio   %sDate   Time   CRC-32    Name\n", y_str);
936 			printf("--------  ------  ------- -----   %s----   ----   ------    ----\n", y_str);
937 		}
938 	}
939 
940 	total_size = 0;
941 	file_count = 0;
942 	error_count = 0;
943 	for (;;) {
944 		ret = archive_read_next_header(a, &e);
945 		if (ret == ARCHIVE_EOF)
946 			break;
947 		ac(ret);
948 		if (!zipinfo_mode) {
949 			if (t_opt)
950 				error_count += test(a, e);
951 			else if (v_opt)
952 				list(a, e);
953 			else if (p_opt || c_opt)
954 				extract_stdout(a, e);
955 			else
956 				extract(a, e);
957 		} else {
958 			if (Z1_opt)
959 				list(a, e);
960 		}
961 
962 		total_size += archive_entry_size(e);
963 		++file_count;
964 	}
965 
966 	if (zipinfo_mode) {
967 		if (v_opt == 1) {
968 			printf(" --------                   %s-------\n", y_str);
969 			printf(" %8ju                   %s%ju file%s\n",
970 			    total_size, y_str, file_count, file_count != 1 ? "s" : "");
971 		} else if (v_opt == 2) {
972 			printf("--------          -------  ---                            %s-------\n", y_str);
973 			printf("%8ju          %7ju   0%%                            %s%ju file%s\n",
974 			    total_size, total_size, y_str, file_count,
975 			    file_count != 1 ? "s" : "");
976 		}
977 	}
978 
979 	ac(archive_read_free(a));
980 
981 	if (passphrase_buf != NULL) {
982 		memset_s(passphrase_buf, PPBUFF_SIZE, 0, PPBUFF_SIZE);
983 		free(passphrase_buf);
984 	}
985 
986 	if (t_opt) {
987 		if (error_count > 0) {
988 			errorx("%ju checksum error(s) found.", error_count);
989 		}
990 		else {
991 			printf("No errors detected in compressed data of %s.\n",
992 			       fn);
993 		}
994 	}
995 }
996 
997 static void
usage(void)998 usage(void)
999 {
1000 
1001 	fprintf(stderr, "Usage: unzip [-aCcfjLlnopqtuvyZ1] [-d dir] "
1002 		"[-x pattern] [-P password] zipfile\n");
1003 	exit(EXIT_FAILURE);
1004 }
1005 
1006 static int
getopts(int argc,char * argv[])1007 getopts(int argc, char *argv[])
1008 {
1009 	int opt;
1010 
1011 	optreset = optind = 1;
1012 	while ((opt = getopt(argc, argv, "aCcd:fjLlnopP:qtuvx:yZ1")) != -1)
1013 		switch (opt) {
1014 		case '1':
1015 			Z1_opt = 1;
1016 			break;
1017 		case 'a':
1018 			a_opt = 1;
1019 			break;
1020 		case 'C':
1021 			C_opt = 1;
1022 			break;
1023 		case 'c':
1024 			c_opt = 1;
1025 			break;
1026 		case 'd':
1027 			d_arg = optarg;
1028 			break;
1029 		case 'f':
1030 			f_opt = 1;
1031 			break;
1032 		case 'j':
1033 			j_opt = 1;
1034 			break;
1035 		case 'L':
1036 			L_opt = 1;
1037 			break;
1038 		case 'l':
1039 			if (v_opt == 0)
1040 				v_opt = 1;
1041 			break;
1042 		case 'n':
1043 			n_opt = 1;
1044 			break;
1045 		case 'o':
1046 			o_opt = 1;
1047 			q_opt = 1;
1048 			break;
1049 		case 'p':
1050 			p_opt = 1;
1051 			break;
1052 		case 'P':
1053 			P_arg = optarg;
1054 			break;
1055 		case 'q':
1056 			q_opt = 1;
1057 			break;
1058 		case 't':
1059 			t_opt = 1;
1060 			break;
1061 		case 'u':
1062 			u_opt = 1;
1063 			break;
1064 		case 'v':
1065 			v_opt = 2;
1066 			break;
1067 		case 'x':
1068 			add_pattern(&exclude, optarg);
1069 			break;
1070 		case 'y':
1071 			y_str = "  ";
1072 			break;
1073 		case 'Z':
1074 			zipinfo_mode = 1;
1075 			break;
1076 		default:
1077 			usage();
1078 		}
1079 
1080 	return (optind);
1081 }
1082 
1083 int
main(int argc,char * argv[])1084 main(int argc, char *argv[])
1085 {
1086 	const char *zipfile;
1087 	int nopts;
1088 
1089 	if (isatty(STDOUT_FILENO))
1090 		tty = 1;
1091 
1092 	if (getenv("UNZIP_DEBUG") != NULL)
1093 		unzip_debug = 1;
1094 	for (int i = 0; i < argc; ++i)
1095 		debug("%s%c", argv[i], (i < argc - 1) ? ' ' : '\n');
1096 
1097 	/*
1098 	 * Info-ZIP's unzip(1) expects certain options to come before the
1099 	 * zipfile name, and others to come after - though it does not
1100 	 * enforce this.  For simplicity, we accept *all* options both
1101 	 * before and after the zipfile name.
1102 	 */
1103 	nopts = getopts(argc, argv);
1104 
1105 	/*
1106 	 * When more of the zipinfo mode options are implemented, this
1107 	 * will need to change.
1108 	 */
1109 	if (zipinfo_mode && !Z1_opt) {
1110 		printf("Zipinfo mode needs additional options\n");
1111 		exit(EXIT_FAILURE);
1112 	}
1113 
1114 	if (argc <= nopts)
1115 		usage();
1116 	zipfile = argv[nopts++];
1117 
1118 	if (strcmp(zipfile, "-") == 0)
1119 		zipfile = NULL; /* STDIN */
1120 
1121 	while (nopts < argc && *argv[nopts] != '-')
1122 		add_pattern(&include, argv[nopts++]);
1123 
1124 	nopts--; /* fake argv[0] */
1125 	nopts += getopts(argc - nopts, argv + nopts);
1126 
1127 	if (n_opt + o_opt + u_opt > 1)
1128 		errorx("-n, -o and -u are contradictory");
1129 
1130 	unzip(zipfile);
1131 
1132 	exit(EXIT_SUCCESS);
1133 }
1134