xref: /freebsd/contrib/libarchive/tar/read.c (revision 537d134373141c2d25bfb24af6d661d0e6102927)
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("$FreeBSD$");
28 
29 #ifdef HAVE_SYS_TYPES_H
30 #include <sys/types.h>
31 #endif
32 #ifdef HAVE_SYS_PARAM_H
33 #include <sys/param.h>
34 #endif
35 #ifdef HAVE_SYS_STAT_H
36 #include <sys/stat.h>
37 #endif
38 
39 #ifdef HAVE_ERRNO_H
40 #include <errno.h>
41 #endif
42 
43 #ifdef HAVE_FCNTL_H
44 #include <fcntl.h>
45 #endif
46 
47 #ifdef HAVE_GRP_H
48 #include <grp.h>
49 #endif
50 
51 #ifdef HAVE_IO_H
52 #include <io.h>
53 #endif
54 
55 #ifdef HAVE_LIMITS_H
56 #include <limits.h>
57 #endif
58 #ifdef HAVE_PWD_H
59 #include <pwd.h>
60 #endif
61 #ifdef HAVE_STDINT_H
62 #include <stdint.h>
63 #endif
64 #include <stdio.h>
65 #ifdef HAVE_STDLIB_H
66 #include <stdlib.h>
67 #endif
68 #ifdef HAVE_STRING_H
69 #include <string.h>
70 #endif
71 #ifdef HAVE_TIME_H
72 #include <time.h>
73 #endif
74 #ifdef HAVE_UNISTD_H
75 #include <unistd.h>
76 #endif
77 
78 #include "bsdtar.h"
79 #include "err.h"
80 
81 struct progress_data {
82 	struct bsdtar *bsdtar;
83 	struct archive *archive;
84 	struct archive_entry *entry;
85 };
86 
87 static void	read_archive(struct bsdtar *bsdtar, char mode, struct archive *);
88 static int unmatched_inclusions_warn(struct archive *matching, const char *);
89 
90 
91 void
92 tar_mode_t(struct bsdtar *bsdtar)
93 {
94 	read_archive(bsdtar, 't', NULL);
95 	if (unmatched_inclusions_warn(bsdtar->matching,
96 	    "Not found in archive") != 0)
97 		bsdtar->return_value = 1;
98 }
99 
100 void
101 tar_mode_x(struct bsdtar *bsdtar)
102 {
103 	struct archive *writer;
104 
105 	writer = archive_write_disk_new();
106 	if (writer == NULL)
107 		lafe_errc(1, ENOMEM, "Cannot allocate disk writer object");
108 	if ((bsdtar->flags & OPTFLAG_NUMERIC_OWNER) == 0)
109 		archive_write_disk_set_standard_lookup(writer);
110 	archive_write_disk_set_options(writer, bsdtar->extract_flags);
111 
112 	read_archive(bsdtar, 'x', writer);
113 
114 	if (unmatched_inclusions_warn(bsdtar->matching,
115 	    "Not found in archive") != 0)
116 		bsdtar->return_value = 1;
117 	archive_write_free(writer);
118 }
119 
120 static void
121 progress_func(void *cookie)
122 {
123 	struct progress_data *progress_data = (struct progress_data *)cookie;
124 	struct bsdtar *bsdtar = progress_data->bsdtar;
125 	struct archive *a = progress_data->archive;
126 	struct archive_entry *entry = progress_data->entry;
127 	uint64_t comp, uncomp;
128 	int compression;
129 
130 	if (!need_report())
131 		return;
132 
133 	if (bsdtar->verbose)
134 		fprintf(stderr, "\n");
135 	if (a != NULL) {
136 		comp = archive_filter_bytes(a, -1);
137 		uncomp = archive_filter_bytes(a, 0);
138 		if (comp > uncomp)
139 			compression = 0;
140 		else
141 			compression = (int)((uncomp - comp) * 100 / uncomp);
142 		fprintf(stderr,
143 		    "In: %s bytes, compression %d%%;",
144 		    tar_i64toa(comp), compression);
145 		fprintf(stderr, "  Out: %d files, %s bytes\n",
146 		    archive_file_count(a), tar_i64toa(uncomp));
147 	}
148 	if (entry != NULL) {
149 		safe_fprintf(stderr, "Current: %s",
150 		    archive_entry_pathname(entry));
151 		fprintf(stderr, " (%s bytes)\n",
152 		    tar_i64toa(archive_entry_size(entry)));
153 	}
154 }
155 
156 /*
157  * Handle 'x' and 't' modes.
158  */
159 static void
160 read_archive(struct bsdtar *bsdtar, char mode, struct archive *writer)
161 {
162 	struct progress_data	progress_data;
163 	FILE			 *out;
164 	struct archive		 *a;
165 	struct archive_entry	 *entry;
166 	const char		 *reader_options;
167 	int			  r;
168 
169 	while (*bsdtar->argv) {
170 		if (archive_match_include_pattern(bsdtar->matching,
171 		    *bsdtar->argv) != ARCHIVE_OK)
172 			lafe_errc(1, 0, "Error inclusion pattern: %s",
173 			    archive_error_string(bsdtar->matching));
174 		bsdtar->argv++;
175 	}
176 
177 	if (bsdtar->names_from_file != NULL)
178 		if (archive_match_include_pattern_from_file(
179 		    bsdtar->matching, bsdtar->names_from_file,
180 		    (bsdtar->flags & OPTFLAG_NULL)) != ARCHIVE_OK)
181 			lafe_errc(1, 0, "Error inclusion pattern: %s",
182 			    archive_error_string(bsdtar->matching));
183 
184 	a = archive_read_new();
185 	if (cset_read_support_filter_program(bsdtar->cset, a) == 0)
186 		archive_read_support_filter_all(a);
187 	archive_read_support_format_all(a);
188 
189 	reader_options = getenv(ENV_READER_OPTIONS);
190 	if (reader_options != NULL) {
191 		size_t module_len = sizeof(IGNORE_WRONG_MODULE_NAME) - 1;
192 		size_t opt_len = strlen(reader_options) + 1;
193 		char *p;
194 		/* Set default read options. */
195 		if ((p = malloc(module_len + opt_len)) == NULL)
196 			lafe_errc(1, errno, "Out of memory");
197 		/* Prepend magic code to ignore options for
198 		 * a format or  modules which are not added to
199 		 *  the archive read object. */
200 		memcpy(p, IGNORE_WRONG_MODULE_NAME, module_len);
201 		memcpy(p + module_len, reader_options, opt_len);
202 		r = archive_read_set_options(a, p);
203 		free(p);
204 		if (r == ARCHIVE_FATAL)
205 			lafe_errc(1, 0, "%s", archive_error_string(a));
206 		else
207 			archive_clear_error(a);
208 	}
209 	if (ARCHIVE_OK != archive_read_set_options(a, bsdtar->option_options))
210 		lafe_errc(1, 0, "%s", archive_error_string(a));
211 	if (bsdtar->flags & OPTFLAG_IGNORE_ZEROS)
212 		if (archive_read_set_options(a,
213 		    "read_concatenated_archives") != ARCHIVE_OK)
214 			lafe_errc(1, 0, "%s", archive_error_string(a));
215 	if (bsdtar->passphrase != NULL)
216 		r = archive_read_add_passphrase(a, bsdtar->passphrase);
217 	else
218 		r = archive_read_set_passphrase_callback(a, bsdtar,
219 			&passphrase_callback);
220 	if (r != ARCHIVE_OK)
221 		lafe_errc(1, 0, "%s", archive_error_string(a));
222 	if (archive_read_open_filename(a, bsdtar->filename,
223 					bsdtar->bytes_per_block))
224 		lafe_errc(1, 0, "Error opening archive: %s",
225 		    archive_error_string(a));
226 
227 	do_chdir(bsdtar);
228 
229 	if (mode == 'x') {
230 		/* Set an extract callback so that we can handle SIGINFO. */
231 		progress_data.bsdtar = bsdtar;
232 		progress_data.archive = a;
233 		archive_read_extract_set_progress_callback(a, progress_func,
234 		    &progress_data);
235 	}
236 
237 	if (mode == 'x' && (bsdtar->flags & OPTFLAG_CHROOT)) {
238 #if HAVE_CHROOT
239 		if (chroot(".") != 0)
240 			lafe_errc(1, errno, "Can't chroot to \".\"");
241 #else
242 		lafe_errc(1, 0,
243 		    "chroot isn't supported on this platform");
244 #endif
245 	}
246 
247 #if defined(_WIN32) && !defined(__CYGWIN__)
248 	if (mode == 'x' && (bsdtar->flags & OPTFLAG_STDOUT)) {
249 		_setmode(1, _O_BINARY);
250 	}
251 #endif
252 
253 	for (;;) {
254 		/* Support --fast-read option */
255 		const char *p;
256 		if ((bsdtar->flags & OPTFLAG_FAST_READ) &&
257 		    archive_match_path_unmatched_inclusions(bsdtar->matching) == 0)
258 			break;
259 
260 		r = archive_read_next_header(a, &entry);
261 		progress_data.entry = entry;
262 		if (r == ARCHIVE_EOF)
263 			break;
264 		if (r < ARCHIVE_OK)
265 			lafe_warnc(0, "%s", archive_error_string(a));
266 		if (r <= ARCHIVE_WARN)
267 			bsdtar->return_value = 1;
268 		if (r == ARCHIVE_RETRY) {
269 			/* Retryable error: try again */
270 			lafe_warnc(0, "Retrying...");
271 			continue;
272 		}
273 		if (r == ARCHIVE_FATAL)
274 			break;
275 		p = archive_entry_pathname(entry);
276 		if (p == NULL || p[0] == '\0') {
277 			lafe_warnc(0, "Archive entry has empty or unreadable filename ... skipping.");
278 			bsdtar->return_value = 1;
279 			continue;
280 		}
281 
282 		if (bsdtar->uid >= 0) {
283 			archive_entry_set_uid(entry, bsdtar->uid);
284 			archive_entry_set_uname(entry, NULL);
285 		}
286 		if (bsdtar->gid >= 0) {
287 			archive_entry_set_gid(entry, bsdtar->gid);
288 			archive_entry_set_gname(entry, NULL);
289 		}
290 		if (bsdtar->uname)
291 			archive_entry_set_uname(entry, bsdtar->uname);
292 		if (bsdtar->gname)
293 			archive_entry_set_gname(entry, bsdtar->gname);
294 
295 		/*
296 		 * Note that pattern exclusions are checked before
297 		 * pathname rewrites are handled.  This gives more
298 		 * control over exclusions, since rewrites always lose
299 		 * information.  (For example, consider a rewrite
300 		 * s/foo[0-9]/foo/.  If we check exclusions after the
301 		 * rewrite, there would be no way to exclude foo1/bar
302 		 * while allowing foo2/bar.)
303 		 */
304 		if (archive_match_excluded(bsdtar->matching, entry))
305 			continue; /* Excluded by a pattern test. */
306 
307 		if (mode == 't') {
308 			/* Perversely, gtar uses -O to mean "send to stderr"
309 			 * when used with -t. */
310 			out = (bsdtar->flags & OPTFLAG_STDOUT) ?
311 			    stderr : stdout;
312 
313 			/*
314 			 * TODO: Provide some reasonable way to
315 			 * preview rewrites.  gtar always displays
316 			 * the unedited path in -t output, which means
317 			 * you cannot easily preview rewrites.
318 			 */
319 			if (bsdtar->verbose < 2)
320 				safe_fprintf(out, "%s",
321 				    archive_entry_pathname(entry));
322 			else
323 				list_item_verbose(bsdtar, out, entry);
324 			fflush(out);
325 			r = archive_read_data_skip(a);
326 			if (r == ARCHIVE_WARN) {
327 				fprintf(out, "\n");
328 				lafe_warnc(0, "%s",
329 				    archive_error_string(a));
330 			}
331 			if (r == ARCHIVE_RETRY) {
332 				fprintf(out, "\n");
333 				lafe_warnc(0, "%s",
334 				    archive_error_string(a));
335 			}
336 			if (r == ARCHIVE_FATAL) {
337 				fprintf(out, "\n");
338 				lafe_warnc(0, "%s",
339 				    archive_error_string(a));
340 				bsdtar->return_value = 1;
341 				break;
342 			}
343 			fprintf(out, "\n");
344 		} else {
345 			/* Note: some rewrite failures prevent extraction. */
346 			if (edit_pathname(bsdtar, entry))
347 				continue; /* Excluded by a rewrite failure. */
348 
349 			if ((bsdtar->flags & OPTFLAG_INTERACTIVE) &&
350 			    !yes("extract '%s'", archive_entry_pathname(entry)))
351 				continue;
352 
353 			if (bsdtar->verbose > 1) {
354 				/* GNU tar uses -tv format with -xvv */
355 				safe_fprintf(stderr, "x ");
356 				list_item_verbose(bsdtar, stderr, entry);
357 				fflush(stderr);
358 			} else if (bsdtar->verbose > 0) {
359 				/* Format follows SUSv2, including the
360 				 * deferred '\n'. */
361 				safe_fprintf(stderr, "x %s",
362 				    archive_entry_pathname(entry));
363 				fflush(stderr);
364 			}
365 
366 			/* TODO siginfo_printinfo(bsdtar, 0); */
367 
368 			if (bsdtar->flags & OPTFLAG_STDOUT)
369 				r = archive_read_data_into_fd(a, 1);
370 			else
371 				r = archive_read_extract2(a, entry, writer);
372 			if (r != ARCHIVE_OK) {
373 				if (!bsdtar->verbose)
374 					safe_fprintf(stderr, "%s",
375 					    archive_entry_pathname(entry));
376 				safe_fprintf(stderr, ": %s",
377 				    archive_error_string(a));
378 				if (!bsdtar->verbose)
379 					fprintf(stderr, "\n");
380 				bsdtar->return_value = 1;
381 			}
382 			if (bsdtar->verbose)
383 				fprintf(stderr, "\n");
384 			if (r == ARCHIVE_FATAL)
385 				break;
386 		}
387 	}
388 
389 
390 	r = archive_read_close(a);
391 	if (r != ARCHIVE_OK)
392 		lafe_warnc(0, "%s", archive_error_string(a));
393 	if (r <= ARCHIVE_WARN)
394 		bsdtar->return_value = 1;
395 
396 	if (bsdtar->verbose > 2)
397 		fprintf(stdout, "Archive Format: %s,  Compression: %s\n",
398 		    archive_format_name(a), archive_filter_name(a, 0));
399 
400 	archive_read_free(a);
401 }
402 
403 
404 static int
405 unmatched_inclusions_warn(struct archive *matching, const char *msg)
406 {
407 	const char *p;
408 	int r;
409 
410 	if (matching == NULL)
411 		return (0);
412 
413 	while ((r = archive_match_path_unmatched_inclusions_next(
414 	    matching, &p)) == ARCHIVE_OK)
415 		lafe_warnc(0, "%s: %s", p, msg);
416 	if (r == ARCHIVE_FATAL)
417 		lafe_errc(1, errno, "Out of memory");
418 
419 	return (archive_match_path_unmatched_inclusions(matching));
420 }
421