xref: /freebsd/contrib/libarchive/tar/read.c (revision 84943d6f38e936ac3b7a3947ca26eeb27a39f938)
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 
28 #ifdef HAVE_SYS_TYPES_H
29 #include <sys/types.h>
30 #endif
31 #ifdef HAVE_SYS_PARAM_H
32 #include <sys/param.h>
33 #endif
34 #ifdef HAVE_SYS_STAT_H
35 #include <sys/stat.h>
36 #endif
37 
38 #ifdef HAVE_ERRNO_H
39 #include <errno.h>
40 #endif
41 
42 #ifdef HAVE_FCNTL_H
43 #include <fcntl.h>
44 #endif
45 
46 #ifdef HAVE_GRP_H
47 #include <grp.h>
48 #endif
49 
50 #ifdef HAVE_IO_H
51 #include <io.h>
52 #endif
53 
54 #ifdef HAVE_LIMITS_H
55 #include <limits.h>
56 #endif
57 #ifdef HAVE_PWD_H
58 #include <pwd.h>
59 #endif
60 #ifdef HAVE_STDINT_H
61 #include <stdint.h>
62 #endif
63 #include <stdio.h>
64 #ifdef HAVE_STDLIB_H
65 #include <stdlib.h>
66 #endif
67 #ifdef HAVE_STRING_H
68 #include <string.h>
69 #endif
70 #ifdef HAVE_TIME_H
71 #include <time.h>
72 #endif
73 #ifdef HAVE_UNISTD_H
74 #include <unistd.h>
75 #endif
76 
77 #include "bsdtar.h"
78 #include "err.h"
79 
80 struct progress_data {
81 	struct bsdtar *bsdtar;
82 	struct archive *archive;
83 	struct archive_entry *entry;
84 };
85 
86 static void	read_archive(struct bsdtar *bsdtar, char mode, struct archive *);
87 static int unmatched_inclusions_warn(struct archive *matching, const char *);
88 
89 
90 void
91 tar_mode_t(struct bsdtar *bsdtar)
92 {
93 	read_archive(bsdtar, 't', NULL);
94 	if (unmatched_inclusions_warn(bsdtar->matching,
95 	    "Not found in archive") != 0)
96 		bsdtar->return_value = 1;
97 }
98 
99 void
100 tar_mode_x(struct bsdtar *bsdtar)
101 {
102 	struct archive *writer;
103 
104 	writer = archive_write_disk_new();
105 	if (writer == NULL)
106 		lafe_errc(1, ENOMEM, "Cannot allocate disk writer object");
107 	if ((bsdtar->flags & OPTFLAG_NUMERIC_OWNER) == 0)
108 		archive_write_disk_set_standard_lookup(writer);
109 	archive_write_disk_set_options(writer, bsdtar->extract_flags);
110 
111 	read_archive(bsdtar, 'x', writer);
112 
113 	if (unmatched_inclusions_warn(bsdtar->matching,
114 	    "Not found in archive") != 0)
115 		bsdtar->return_value = 1;
116 	archive_write_free(writer);
117 }
118 
119 static void
120 progress_func(void *cookie)
121 {
122 	struct progress_data *progress_data = (struct progress_data *)cookie;
123 	struct bsdtar *bsdtar = progress_data->bsdtar;
124 	struct archive *a = progress_data->archive;
125 	struct archive_entry *entry = progress_data->entry;
126 	uint64_t comp, uncomp;
127 	int compression;
128 
129 	if (!need_report())
130 		return;
131 
132 	if (bsdtar->verbose)
133 		fprintf(stderr, "\n");
134 	if (a != NULL) {
135 		comp = archive_filter_bytes(a, -1);
136 		uncomp = archive_filter_bytes(a, 0);
137 		if (comp > uncomp)
138 			compression = 0;
139 		else
140 			compression = (int)((uncomp - comp) * 100 / uncomp);
141 		fprintf(stderr,
142 		    "In: %s bytes, compression %d%%;",
143 		    tar_i64toa(comp), compression);
144 		fprintf(stderr, "  Out: %d files, %s bytes\n",
145 		    archive_file_count(a), tar_i64toa(uncomp));
146 	}
147 	if (entry != NULL) {
148 		safe_fprintf(stderr, "Current: %s",
149 		    archive_entry_pathname(entry));
150 		fprintf(stderr, " (%s bytes)\n",
151 		    tar_i64toa(archive_entry_size(entry)));
152 	}
153 }
154 
155 /*
156  * Handle 'x' and 't' modes.
157  */
158 static void
159 read_archive(struct bsdtar *bsdtar, char mode, struct archive *writer)
160 {
161 	struct progress_data	progress_data;
162 	FILE			 *out;
163 	struct archive		 *a;
164 	struct archive_entry	 *entry;
165 	const char		 *reader_options;
166 	int			  r;
167 
168 	while (*bsdtar->argv) {
169 		if (archive_match_include_pattern(bsdtar->matching,
170 		    *bsdtar->argv) != ARCHIVE_OK)
171 			lafe_errc(1, 0, "Error inclusion pattern: %s",
172 			    archive_error_string(bsdtar->matching));
173 		bsdtar->argv++;
174 	}
175 
176 	if (bsdtar->names_from_file != NULL)
177 		if (archive_match_include_pattern_from_file(
178 		    bsdtar->matching, bsdtar->names_from_file,
179 		    (bsdtar->flags & OPTFLAG_NULL)) != ARCHIVE_OK)
180 			lafe_errc(1, 0, "Error inclusion pattern: %s",
181 			    archive_error_string(bsdtar->matching));
182 
183 	a = archive_read_new();
184 	if (cset_read_support_filter_program(bsdtar->cset, a) == 0)
185 		archive_read_support_filter_all(a);
186 	archive_read_support_format_all(a);
187 
188 	reader_options = getenv(ENV_READER_OPTIONS);
189 	if (reader_options != NULL) {
190 		size_t module_len = sizeof(IGNORE_WRONG_MODULE_NAME) - 1;
191 		size_t opt_len = strlen(reader_options) + 1;
192 		char *p;
193 		/* Set default read options. */
194 		if ((p = malloc(module_len + opt_len)) == NULL)
195 			lafe_errc(1, errno, "Out of memory");
196 		/* Prepend magic code to ignore options for
197 		 * a format or  modules which are not added to
198 		 *  the archive read object. */
199 		memcpy(p, IGNORE_WRONG_MODULE_NAME, module_len);
200 		memcpy(p + module_len, reader_options, opt_len);
201 		r = archive_read_set_options(a, p);
202 		free(p);
203 		if (r == ARCHIVE_FATAL)
204 			lafe_errc(1, 0, "%s", archive_error_string(a));
205 		else
206 			archive_clear_error(a);
207 	}
208 	if (ARCHIVE_OK != archive_read_set_options(a, bsdtar->option_options))
209 		lafe_errc(1, 0, "%s", archive_error_string(a));
210 	if (bsdtar->flags & OPTFLAG_IGNORE_ZEROS)
211 		if (archive_read_set_options(a,
212 		    "read_concatenated_archives") != ARCHIVE_OK)
213 			lafe_errc(1, 0, "%s", archive_error_string(a));
214 	if (bsdtar->passphrase != NULL)
215 		r = archive_read_add_passphrase(a, bsdtar->passphrase);
216 	else
217 		r = archive_read_set_passphrase_callback(a, bsdtar,
218 			&passphrase_callback);
219 	if (r != ARCHIVE_OK)
220 		lafe_errc(1, 0, "%s", archive_error_string(a));
221 	if (archive_read_open_filename(a, bsdtar->filename,
222 					bsdtar->bytes_per_block))
223 		lafe_errc(1, 0, "Error opening archive: %s",
224 		    archive_error_string(a));
225 
226 	do_chdir(bsdtar);
227 
228 	if (mode == 'x') {
229 		/* Set an extract callback so that we can handle SIGINFO. */
230 		progress_data.bsdtar = bsdtar;
231 		progress_data.archive = a;
232 		archive_read_extract_set_progress_callback(a, progress_func,
233 		    &progress_data);
234 	}
235 
236 	if (mode == 'x' && (bsdtar->flags & OPTFLAG_CHROOT)) {
237 #if HAVE_CHROOT
238 		if (chroot(".") != 0)
239 			lafe_errc(1, errno, "Can't chroot to \".\"");
240 #else
241 		lafe_errc(1, 0,
242 		    "chroot isn't supported on this platform");
243 #endif
244 	}
245 
246 #if defined(_WIN32) && !defined(__CYGWIN__)
247 	if (mode == 'x' && (bsdtar->flags & OPTFLAG_STDOUT)) {
248 		_setmode(1, _O_BINARY);
249 	}
250 #endif
251 
252 	for (;;) {
253 		/* Support --fast-read option */
254 		const char *p;
255 		if ((bsdtar->flags & OPTFLAG_FAST_READ) &&
256 		    archive_match_path_unmatched_inclusions(bsdtar->matching) == 0)
257 			break;
258 
259 		r = archive_read_next_header(a, &entry);
260 		progress_data.entry = entry;
261 		if (r == ARCHIVE_EOF)
262 			break;
263 		if (r < ARCHIVE_OK)
264 			lafe_warnc(0, "%s", archive_error_string(a));
265 		if (r <= ARCHIVE_WARN)
266 			bsdtar->return_value = 1;
267 		if (r == ARCHIVE_RETRY) {
268 			/* Retryable error: try again */
269 			lafe_warnc(0, "Retrying...");
270 			continue;
271 		}
272 		if (r == ARCHIVE_FATAL)
273 			break;
274 		p = archive_entry_pathname(entry);
275 		if (p == NULL || p[0] == '\0') {
276 			lafe_warnc(0, "Archive entry has empty or unreadable filename ... skipping.");
277 			bsdtar->return_value = 1;
278 			continue;
279 		}
280 
281 		if (bsdtar->uid >= 0) {
282 			archive_entry_set_uid(entry, bsdtar->uid);
283 			archive_entry_set_uname(entry, NULL);
284 		}
285 		if (bsdtar->gid >= 0) {
286 			archive_entry_set_gid(entry, bsdtar->gid);
287 			archive_entry_set_gname(entry, NULL);
288 		}
289 		if (bsdtar->uname)
290 			archive_entry_set_uname(entry, bsdtar->uname);
291 		if (bsdtar->gname)
292 			archive_entry_set_gname(entry, bsdtar->gname);
293 
294 		/*
295 		 * Note that pattern exclusions are checked before
296 		 * pathname rewrites are handled.  This gives more
297 		 * control over exclusions, since rewrites always lose
298 		 * information.  (For example, consider a rewrite
299 		 * s/foo[0-9]/foo/.  If we check exclusions after the
300 		 * rewrite, there would be no way to exclude foo1/bar
301 		 * while allowing foo2/bar.)
302 		 */
303 		if (archive_match_excluded(bsdtar->matching, entry))
304 			continue; /* Excluded by a pattern test. */
305 
306 		if (mode == 't') {
307 			/* Perversely, gtar uses -O to mean "send to stderr"
308 			 * when used with -t. */
309 			out = (bsdtar->flags & OPTFLAG_STDOUT) ?
310 			    stderr : stdout;
311 
312 			/*
313 			 * TODO: Provide some reasonable way to
314 			 * preview rewrites.  gtar always displays
315 			 * the unedited path in -t output, which means
316 			 * you cannot easily preview rewrites.
317 			 */
318 			if (bsdtar->verbose < 2)
319 				safe_fprintf(out, "%s",
320 				    archive_entry_pathname(entry));
321 			else
322 				list_item_verbose(bsdtar, out, entry);
323 			fflush(out);
324 			r = archive_read_data_skip(a);
325 			if (r == ARCHIVE_WARN) {
326 				fprintf(out, "\n");
327 				lafe_warnc(0, "%s",
328 				    archive_error_string(a));
329 			}
330 			if (r == ARCHIVE_RETRY) {
331 				fprintf(out, "\n");
332 				lafe_warnc(0, "%s",
333 				    archive_error_string(a));
334 			}
335 			if (r == ARCHIVE_FATAL) {
336 				fprintf(out, "\n");
337 				lafe_warnc(0, "%s",
338 				    archive_error_string(a));
339 				bsdtar->return_value = 1;
340 				break;
341 			}
342 			fprintf(out, "\n");
343 		} else {
344 			/* Note: some rewrite failures prevent extraction. */
345 			if (edit_pathname(bsdtar, entry))
346 				continue; /* Excluded by a rewrite failure. */
347 
348 			if ((bsdtar->flags & OPTFLAG_INTERACTIVE) &&
349 			    !yes("extract '%s'", archive_entry_pathname(entry)))
350 				continue;
351 
352 			if (bsdtar->verbose > 1) {
353 				/* GNU tar uses -tv format with -xvv */
354 				safe_fprintf(stderr, "x ");
355 				list_item_verbose(bsdtar, stderr, entry);
356 				fflush(stderr);
357 			} else if (bsdtar->verbose > 0) {
358 				/* Format follows SUSv2, including the
359 				 * deferred '\n'. */
360 				safe_fprintf(stderr, "x %s",
361 				    archive_entry_pathname(entry));
362 				fflush(stderr);
363 			}
364 
365 			/* TODO siginfo_printinfo(bsdtar, 0); */
366 
367 			if (bsdtar->flags & OPTFLAG_STDOUT)
368 				r = archive_read_data_into_fd(a, 1);
369 			else
370 				r = archive_read_extract2(a, entry, writer);
371 			if (r != ARCHIVE_OK) {
372 				if (!bsdtar->verbose)
373 					safe_fprintf(stderr, "%s", archive_entry_pathname(entry));
374 				safe_fprintf(stderr, ": %s: %s",
375 				    archive_error_string(a),
376 				    strerror(archive_errno(a)));
377 				if (!bsdtar->verbose)
378 					fprintf(stderr, "\n");
379 				bsdtar->return_value = 1;
380 			}
381 			if (bsdtar->verbose)
382 				fprintf(stderr, "\n");
383 			if (r == ARCHIVE_FATAL)
384 				break;
385 		}
386 	}
387 
388 
389 	r = archive_read_close(a);
390 	if (r != ARCHIVE_OK)
391 		lafe_warnc(0, "%s", archive_error_string(a));
392 	if (r <= ARCHIVE_WARN)
393 		bsdtar->return_value = 1;
394 
395 	if (bsdtar->verbose > 2)
396 		fprintf(stdout, "Archive Format: %s,  Compression: %s\n",
397 		    archive_format_name(a), archive_filter_name(a, 0));
398 
399 	archive_read_free(a);
400 }
401 
402 
403 static int
404 unmatched_inclusions_warn(struct archive *matching, const char *msg)
405 {
406 	const char *p;
407 	int r;
408 
409 	if (matching == NULL)
410 		return (0);
411 
412 	while ((r = archive_match_path_unmatched_inclusions_next(
413 	    matching, &p)) == ARCHIVE_OK)
414 		lafe_warnc(0, "%s: %s", p, msg);
415 	if (r == ARCHIVE_FATAL)
416 		lafe_errc(1, errno, "Out of memory");
417 
418 	return (archive_match_path_unmatched_inclusions(matching));
419 }
420