xref: /dragonfly/contrib/libarchive/tar/read.c (revision 9f3fc534)
1 /*-
2  * Copyright (c) 2003-2007 Tim Kientzle
3  * All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice, this list of conditions and the following disclaimer.
10  * 2. Redistributions in binary form must reproduce the above copyright
11  *    notice, this list of conditions and the following disclaimer in the
12  *    documentation and/or other materials provided with the distribution.
13  *
14  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
15  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
16  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
17  * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
18  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
19  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
20  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
21  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
22  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
23  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
24  */
25 
26 #include "bsdtar_platform.h"
27 __FBSDID("$FreeBSD: src/usr.bin/tar/read.c,v 1.38 2008/05/26 17:10:10 kientzle Exp $");
28 
29 #ifdef HAVE_SYS_TYPES_H
30 #include <sys/types.h>
31 #endif
32 #ifdef MAJOR_IN_MKDEV
33 #include <sys/mkdev.h>
34 #elif defined(MAJOR_IN_SYSMACROS)
35 #include <sys/sysmacros.h>
36 #endif
37 #ifdef HAVE_SYS_PARAM_H
38 #include <sys/param.h>
39 #endif
40 #ifdef HAVE_SYS_STAT_H
41 #include <sys/stat.h>
42 #endif
43 
44 #ifdef HAVE_ERRNO_H
45 #include <errno.h>
46 #endif
47 #ifdef HAVE_GRP_H
48 #include <grp.h>
49 #endif
50 #ifdef HAVE_LIMITS_H
51 #include <limits.h>
52 #endif
53 #ifdef HAVE_PWD_H
54 #include <pwd.h>
55 #endif
56 #include <stdio.h>
57 #ifdef HAVE_STDLIB_H
58 #include <stdlib.h>
59 #endif
60 #ifdef HAVE_STRING_H
61 #include <string.h>
62 #endif
63 #ifdef HAVE_TIME_H
64 #include <time.h>
65 #endif
66 #ifdef HAVE_UNISTD_H
67 #include <unistd.h>
68 #endif
69 
70 #include "bsdtar.h"
71 
72 static void	list_item_verbose(struct bsdtar *, FILE *,
73 		    struct archive_entry *);
74 static void	read_archive(struct bsdtar *bsdtar, char mode);
75 
76 void
77 tar_mode_t(struct bsdtar *bsdtar)
78 {
79 	read_archive(bsdtar, 't');
80 	unmatched_inclusions_warn(bsdtar, "Not found in archive");
81 }
82 
83 void
84 tar_mode_x(struct bsdtar *bsdtar)
85 {
86 	/* We want to catch SIGINFO and SIGUSR1. */
87 	siginfo_init(bsdtar);
88 
89 	read_archive(bsdtar, 'x');
90 
91 	unmatched_inclusions_warn(bsdtar, "Not found in archive");
92 	/* Restore old SIGINFO + SIGUSR1 handlers. */
93 	siginfo_done(bsdtar);
94 }
95 
96 static void
97 progress_func(void * cookie)
98 {
99 	struct bsdtar * bsdtar = cookie;
100 
101 	siginfo_printinfo(bsdtar, 0);
102 }
103 
104 /*
105  * Handle 'x' and 't' modes.
106  */
107 static void
108 read_archive(struct bsdtar *bsdtar, char mode)
109 {
110 	FILE			 *out;
111 	struct archive		 *a;
112 	struct archive_entry	 *entry;
113 	const struct stat	 *st;
114 	int			  r;
115 
116 	while (*bsdtar->argv) {
117 		include(bsdtar, *bsdtar->argv);
118 		bsdtar->argv++;
119 	}
120 
121 	if (bsdtar->names_from_file != NULL)
122 		include_from_file(bsdtar, bsdtar->names_from_file);
123 
124 	a = archive_read_new();
125 	if (bsdtar->compress_program != NULL)
126 		archive_read_support_compression_program(a, bsdtar->compress_program);
127 	else
128 		archive_read_support_compression_all(a);
129 	archive_read_support_format_all(a);
130 	if (archive_read_open_file(a, bsdtar->filename,
131 	    bsdtar->bytes_per_block != 0 ? bsdtar->bytes_per_block :
132 	    DEFAULT_BYTES_PER_BLOCK))
133 		bsdtar_errc(bsdtar, 1, 0, "Error opening archive: %s",
134 		    archive_error_string(a));
135 
136 	do_chdir(bsdtar);
137 
138 	if (mode == 'x') {
139 		/* Set an extract callback so that we can handle SIGINFO. */
140 		archive_read_extract_set_progress_callback(a, progress_func,
141 		    bsdtar);
142 	}
143 
144 	if (mode == 'x' && bsdtar->option_chroot) {
145 #if HAVE_CHROOT
146 		if (chroot(".") != 0)
147 			bsdtar_errc(bsdtar, 1, errno, "Can't chroot to \".\"");
148 #else
149 		bsdtar_errc(bsdtar, 1, 0,
150 		    "chroot isn't supported on this platform");
151 #endif
152 	}
153 
154 	for (;;) {
155 		/* Support --fast-read option */
156 		if (bsdtar->option_fast_read &&
157 		    unmatched_inclusions(bsdtar) == 0)
158 			break;
159 
160 		r = archive_read_next_header(a, &entry);
161 		if (r == ARCHIVE_EOF)
162 			break;
163 		if (r < ARCHIVE_OK)
164 			bsdtar_warnc(bsdtar, 0, "%s", archive_error_string(a));
165 		if (r <= ARCHIVE_WARN)
166 			bsdtar->return_value = 1;
167 		if (r == ARCHIVE_RETRY) {
168 			/* Retryable error: try again */
169 			bsdtar_warnc(bsdtar, 0, "Retrying...");
170 			continue;
171 		}
172 		if (r == ARCHIVE_FATAL)
173 			break;
174 
175 		if (bsdtar->option_numeric_owner) {
176 			archive_entry_set_uname(entry, NULL);
177 			archive_entry_set_gname(entry, NULL);
178 		}
179 
180 		/*
181 		 * Exclude entries that are too old.
182 		 */
183 		st = archive_entry_stat(entry);
184 		if (bsdtar->newer_ctime_sec > 0) {
185 			if (st->st_ctime < bsdtar->newer_ctime_sec)
186 				continue; /* Too old, skip it. */
187 			if (st->st_ctime == bsdtar->newer_ctime_sec
188 			    && ARCHIVE_STAT_CTIME_NANOS(st)
189 			    <= bsdtar->newer_ctime_nsec)
190 				continue; /* Too old, skip it. */
191 		}
192 		if (bsdtar->newer_mtime_sec > 0) {
193 			if (st->st_mtime < bsdtar->newer_mtime_sec)
194 				continue; /* Too old, skip it. */
195 			if (st->st_mtime == bsdtar->newer_mtime_sec
196 			    && ARCHIVE_STAT_MTIME_NANOS(st)
197 			    <= bsdtar->newer_mtime_nsec)
198 				continue; /* Too old, skip it. */
199 		}
200 
201 		/*
202 		 * Note that pattern exclusions are checked before
203 		 * pathname rewrites are handled.  This gives more
204 		 * control over exclusions, since rewrites always lose
205 		 * information.  (For example, consider a rewrite
206 		 * s/foo[0-9]/foo/.  If we check exclusions after the
207 		 * rewrite, there would be no way to exclude foo1/bar
208 		 * while allowing foo2/bar.)
209 		 */
210 		if (excluded(bsdtar, archive_entry_pathname(entry)))
211 			continue; /* Excluded by a pattern test. */
212 
213 		/*
214 		 * Modify the pathname as requested by the user.  We
215 		 * do this for -t as well to give users a way to
216 		 * preview the effects of their rewrites.  We also do
217 		 * this before extraction security checks (including
218 		 * leading '/' removal).  Note that some rewrite
219 		 * failures prevent extraction.
220 		 */
221 		if (edit_pathname(bsdtar, entry))
222 			continue; /* Excluded by a rewrite failure. */
223 
224 		if (mode == 't') {
225 			/* Perversely, gtar uses -O to mean "send to stderr"
226 			 * when used with -t. */
227 			out = bsdtar->option_stdout ? stderr : stdout;
228 
229 			if (bsdtar->verbose < 2)
230 				safe_fprintf(out, "%s",
231 				    archive_entry_pathname(entry));
232 			else
233 				list_item_verbose(bsdtar, out, entry);
234 			fflush(out);
235 			r = archive_read_data_skip(a);
236 			if (r == ARCHIVE_WARN) {
237 				fprintf(out, "\n");
238 				bsdtar_warnc(bsdtar, 0, "%s",
239 				    archive_error_string(a));
240 			}
241 			if (r == ARCHIVE_RETRY) {
242 				fprintf(out, "\n");
243 				bsdtar_warnc(bsdtar, 0, "%s",
244 				    archive_error_string(a));
245 			}
246 			if (r == ARCHIVE_FATAL) {
247 				fprintf(out, "\n");
248 				bsdtar_warnc(bsdtar, 0, "%s",
249 				    archive_error_string(a));
250 				bsdtar->return_value = 1;
251 				break;
252 			}
253 			fprintf(out, "\n");
254 		} else {
255 			if (bsdtar->option_interactive &&
256 			    !yes("extract '%s'", archive_entry_pathname(entry)))
257 				continue;
258 
259 			/*
260 			 * Format here is from SUSv2, including the
261 			 * deferred '\n'.
262 			 */
263 			if (bsdtar->verbose) {
264 				safe_fprintf(stderr, "x %s",
265 				    archive_entry_pathname(entry));
266 				fflush(stderr);
267 			}
268 
269 			/* Tell the SIGINFO-handler code what we're doing. */
270 			siginfo_setinfo(bsdtar, "extracting",
271 			    archive_entry_pathname(entry), 0);
272 			siginfo_printinfo(bsdtar, 0);
273 
274 			if (bsdtar->option_stdout)
275 				r = archive_read_data_into_fd(a, 1);
276 			else
277 				r = archive_read_extract(a, entry,
278 				    bsdtar->extract_flags);
279 			if (r != ARCHIVE_OK) {
280 				if (!bsdtar->verbose)
281 					safe_fprintf(stderr, "%s",
282 					    archive_entry_pathname(entry));
283 				safe_fprintf(stderr, ": %s",
284 				    archive_error_string(a));
285 				if (!bsdtar->verbose)
286 					fprintf(stderr, "\n");
287 				bsdtar->return_value = 1;
288 			}
289 			if (bsdtar->verbose)
290 				fprintf(stderr, "\n");
291 			if (r == ARCHIVE_FATAL)
292 				break;
293 		}
294 	}
295 
296 	if (bsdtar->verbose > 2)
297 		fprintf(stdout, "Archive Format: %s,  Compression: %s\n",
298 		    archive_format_name(a), archive_compression_name(a));
299 
300 	archive_read_finish(a);
301 }
302 
303 
304 /*
305  * Display information about the current file.
306  *
307  * The format here roughly duplicates the output of 'ls -l'.
308  * This is based on SUSv2, where 'tar tv' is documented as
309  * listing additional information in an "unspecified format,"
310  * and 'pax -l' is documented as using the same format as 'ls -l'.
311  */
312 static void
313 list_item_verbose(struct bsdtar *bsdtar, FILE *out, struct archive_entry *entry)
314 {
315 	const struct stat	*st;
316 	char			 tmp[100];
317 	size_t			 w;
318 	const char		*p;
319 	const char		*fmt;
320 	time_t			 tim;
321 	static time_t		 now;
322 
323 	st = archive_entry_stat(entry);
324 
325 	/*
326 	 * We avoid collecting the entire list in memory at once by
327 	 * listing things as we see them.  However, that also means we can't
328 	 * just pre-compute the field widths.  Instead, we start with guesses
329 	 * and just widen them as necessary.  These numbers are completely
330 	 * arbitrary.
331 	 */
332 	if (!bsdtar->u_width) {
333 		bsdtar->u_width = 6;
334 		bsdtar->gs_width = 13;
335 	}
336 	if (!now)
337 		time(&now);
338 	fprintf(out, "%s %d ",
339 	    archive_entry_strmode(entry),
340 	    (int)(st->st_nlink));
341 
342 	/* Use uname if it's present, else uid. */
343 	p = archive_entry_uname(entry);
344 	if ((p == NULL) || (*p == '\0')) {
345 		sprintf(tmp, "%lu ", (unsigned long)st->st_uid);
346 		p = tmp;
347 	}
348 	w = strlen(p);
349 	if (w > bsdtar->u_width)
350 		bsdtar->u_width = w;
351 	fprintf(out, "%-*s ", (int)bsdtar->u_width, p);
352 
353 	/* Use gname if it's present, else gid. */
354 	p = archive_entry_gname(entry);
355 	if (p != NULL && p[0] != '\0') {
356 		fprintf(out, "%s", p);
357 		w = strlen(p);
358 	} else {
359 		sprintf(tmp, "%lu", (unsigned long)st->st_gid);
360 		w = strlen(tmp);
361 		fprintf(out, "%s", tmp);
362 	}
363 
364 	/*
365 	 * Print device number or file size, right-aligned so as to make
366 	 * total width of group and devnum/filesize fields be gs_width.
367 	 * If gs_width is too small, grow it.
368 	 */
369 	if (S_ISCHR(st->st_mode) || S_ISBLK(st->st_mode)) {
370 		sprintf(tmp, "%lu,%lu",
371 		    (unsigned long)major(st->st_rdev),
372 		    (unsigned long)minor(st->st_rdev)); /* ls(1) also casts here. */
373 	} else {
374 		/*
375 		 * Note the use of platform-dependent macros to format
376 		 * the filesize here.  We need the format string and the
377 		 * corresponding type for the cast.
378 		 */
379 		sprintf(tmp, BSDTAR_FILESIZE_PRINTF,
380 		    (BSDTAR_FILESIZE_TYPE)st->st_size);
381 	}
382 	if (w + strlen(tmp) >= bsdtar->gs_width)
383 		bsdtar->gs_width = w+strlen(tmp)+1;
384 	fprintf(out, "%*s", (int)(bsdtar->gs_width - w), tmp);
385 
386 	/* Format the time using 'ls -l' conventions. */
387 	tim = (time_t)st->st_mtime;
388 	if (abs(tim - now) > (365/2)*86400)
389 		fmt = bsdtar->day_first ? "%e %b  %Y" : "%b %e  %Y";
390 	else
391 		fmt = bsdtar->day_first ? "%e %b %H:%M" : "%b %e %H:%M";
392 	strftime(tmp, sizeof(tmp), fmt, localtime(&tim));
393 	fprintf(out, " %s ", tmp);
394 	safe_fprintf(out, "%s", archive_entry_pathname(entry));
395 
396 	/* Extra information for links. */
397 	if (archive_entry_hardlink(entry)) /* Hard link */
398 		safe_fprintf(out, " link to %s",
399 		    archive_entry_hardlink(entry));
400 	else if (S_ISLNK(st->st_mode)) /* Symbolic link */
401 		safe_fprintf(out, " -> %s", archive_entry_symlink(entry));
402 }
403