xref: /freebsd/contrib/libarchive/tar/read.c (revision 3e65b9c6e6b7b2081d54e1dc40983c3c00eaf738)
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 #ifdef HAVE_GRP_H
43 #include <grp.h>
44 #endif
45 #ifdef HAVE_LIMITS_H
46 #include <limits.h>
47 #endif
48 #ifdef HAVE_PWD_H
49 #include <pwd.h>
50 #endif
51 #ifdef HAVE_STDINT_H
52 #include <stdint.h>
53 #endif
54 #include <stdio.h>
55 #ifdef HAVE_STDLIB_H
56 #include <stdlib.h>
57 #endif
58 #ifdef HAVE_STRING_H
59 #include <string.h>
60 #endif
61 #ifdef HAVE_TIME_H
62 #include <time.h>
63 #endif
64 #ifdef HAVE_UNISTD_H
65 #include <unistd.h>
66 #endif
67 
68 #include "bsdtar.h"
69 #include "err.h"
70 
71 struct progress_data {
72 	struct bsdtar *bsdtar;
73 	struct archive *archive;
74 	struct archive_entry *entry;
75 };
76 
77 static void	list_item_verbose(struct bsdtar *, FILE *,
78 		    struct archive_entry *);
79 static void	read_archive(struct bsdtar *bsdtar, char mode);
80 
81 void
82 tar_mode_t(struct bsdtar *bsdtar)
83 {
84 	read_archive(bsdtar, 't');
85 	if (lafe_unmatched_inclusions_warn(bsdtar->matching, "Not found in archive") != 0)
86 		bsdtar->return_value = 1;
87 }
88 
89 void
90 tar_mode_x(struct bsdtar *bsdtar)
91 {
92 	read_archive(bsdtar, 'x');
93 
94 	if (lafe_unmatched_inclusions_warn(bsdtar->matching, "Not found in archive") != 0)
95 		bsdtar->return_value = 1;
96 }
97 
98 static void
99 progress_func(void *cookie)
100 {
101 	struct progress_data *progress_data = cookie;
102 	struct bsdtar *bsdtar = progress_data->bsdtar;
103 	struct archive *a = progress_data->archive;
104 	struct archive_entry *entry = progress_data->entry;
105 	uint64_t comp, uncomp;
106 	int compression;
107 
108 	if (!need_report())
109 		return;
110 
111 	if (bsdtar->verbose)
112 		fprintf(stderr, "\n");
113 	if (a != NULL) {
114 		comp = archive_position_compressed(a);
115 		uncomp = archive_position_uncompressed(a);
116 		if (comp > uncomp)
117 			compression = 0;
118 		else
119 			compression = (int)((uncomp - comp) * 100 / uncomp);
120 		fprintf(stderr,
121 		    "In: %s bytes, compression %d%%;",
122 		    tar_i64toa(comp), compression);
123 		fprintf(stderr, "  Out: %d files, %s bytes\n",
124 		    archive_file_count(a), tar_i64toa(uncomp));
125 	}
126 	if (entry != NULL) {
127 		safe_fprintf(stderr, "Current: %s",
128 		    archive_entry_pathname(entry));
129 		fprintf(stderr, " (%s bytes)\n",
130 		    tar_i64toa(archive_entry_size(entry)));
131 	}
132 }
133 
134 /*
135  * Handle 'x' and 't' modes.
136  */
137 static void
138 read_archive(struct bsdtar *bsdtar, char mode)
139 {
140 	struct progress_data	progress_data;
141 	FILE			 *out;
142 	struct archive		 *a;
143 	struct archive_entry	 *entry;
144 	const struct stat	 *st;
145 	int			  r;
146 
147 	while (*bsdtar->argv) {
148 		lafe_include(&bsdtar->matching, *bsdtar->argv);
149 		bsdtar->argv++;
150 	}
151 
152 	if (bsdtar->names_from_file != NULL)
153 		lafe_include_from_file(&bsdtar->matching,
154 		    bsdtar->names_from_file, bsdtar->option_null);
155 
156 	a = archive_read_new();
157 	if (bsdtar->compress_program != NULL)
158 		archive_read_support_compression_program(a, bsdtar->compress_program);
159 	else
160 		archive_read_support_compression_all(a);
161 	archive_read_support_format_all(a);
162 	if (ARCHIVE_OK != archive_read_set_options(a, bsdtar->option_options))
163 		lafe_errc(1, 0, "%s", archive_error_string(a));
164 	if (archive_read_open_file(a, bsdtar->filename,
165 	    bsdtar->bytes_per_block != 0 ? bsdtar->bytes_per_block :
166 	    DEFAULT_BYTES_PER_BLOCK))
167 		lafe_errc(1, 0, "Error opening archive: %s",
168 		    archive_error_string(a));
169 
170 	do_chdir(bsdtar);
171 
172 	if (mode == 'x') {
173 		/* Set an extract callback so that we can handle SIGINFO. */
174 		progress_data.bsdtar = bsdtar;
175 		progress_data.archive = a;
176 		archive_read_extract_set_progress_callback(a, progress_func,
177 		    &progress_data);
178 	}
179 
180 	if (mode == 'x' && bsdtar->option_chroot) {
181 #if HAVE_CHROOT
182 		if (chroot(".") != 0)
183 			lafe_errc(1, errno, "Can't chroot to \".\"");
184 #else
185 		lafe_errc(1, 0,
186 		    "chroot isn't supported on this platform");
187 #endif
188 	}
189 
190 	for (;;) {
191 		/* Support --fast-read option */
192 		if (bsdtar->option_fast_read &&
193 		    lafe_unmatched_inclusions(bsdtar->matching) == 0)
194 			break;
195 
196 		r = archive_read_next_header(a, &entry);
197 		progress_data.entry = entry;
198 		if (r == ARCHIVE_EOF)
199 			break;
200 		if (r < ARCHIVE_OK)
201 			lafe_warnc(0, "%s", archive_error_string(a));
202 		if (r <= ARCHIVE_WARN)
203 			bsdtar->return_value = 1;
204 		if (r == ARCHIVE_RETRY) {
205 			/* Retryable error: try again */
206 			lafe_warnc(0, "Retrying...");
207 			continue;
208 		}
209 		if (r == ARCHIVE_FATAL)
210 			break;
211 
212 		if (bsdtar->uid >= 0) {
213 			archive_entry_set_uid(entry, bsdtar->uid);
214 			archive_entry_set_uname(entry, NULL);
215 		}
216 		if (bsdtar->gid >= 0) {
217 			archive_entry_set_gid(entry, bsdtar->gid);
218 			archive_entry_set_gname(entry, NULL);
219 		}
220 		if (bsdtar->uname)
221 			archive_entry_set_uname(entry, bsdtar->uname);
222 		if (bsdtar->gname)
223 			archive_entry_set_gname(entry, bsdtar->gname);
224 
225 		/*
226 		 * Exclude entries that are too old.
227 		 */
228 		st = archive_entry_stat(entry);
229 		if (bsdtar->newer_ctime_sec > 0) {
230 			if (st->st_ctime < bsdtar->newer_ctime_sec)
231 				continue; /* Too old, skip it. */
232 			if (st->st_ctime == bsdtar->newer_ctime_sec
233 			    && ARCHIVE_STAT_CTIME_NANOS(st)
234 			    <= bsdtar->newer_ctime_nsec)
235 				continue; /* Too old, skip it. */
236 		}
237 		if (bsdtar->newer_mtime_sec > 0) {
238 			if (st->st_mtime < bsdtar->newer_mtime_sec)
239 				continue; /* Too old, skip it. */
240 			if (st->st_mtime == bsdtar->newer_mtime_sec
241 			    && ARCHIVE_STAT_MTIME_NANOS(st)
242 			    <= bsdtar->newer_mtime_nsec)
243 				continue; /* Too old, skip it. */
244 		}
245 
246 		/*
247 		 * Note that pattern exclusions are checked before
248 		 * pathname rewrites are handled.  This gives more
249 		 * control over exclusions, since rewrites always lose
250 		 * information.  (For example, consider a rewrite
251 		 * s/foo[0-9]/foo/.  If we check exclusions after the
252 		 * rewrite, there would be no way to exclude foo1/bar
253 		 * while allowing foo2/bar.)
254 		 */
255 		if (lafe_excluded(bsdtar->matching, archive_entry_pathname(entry)))
256 			continue; /* Excluded by a pattern test. */
257 
258 		if (mode == 't') {
259 			/* Perversely, gtar uses -O to mean "send to stderr"
260 			 * when used with -t. */
261 			out = bsdtar->option_stdout ? stderr : stdout;
262 
263 			/*
264 			 * TODO: Provide some reasonable way to
265 			 * preview rewrites.  gtar always displays
266 			 * the unedited path in -t output, which means
267 			 * you cannot easily preview rewrites.
268 			 */
269 			if (bsdtar->verbose < 2)
270 				safe_fprintf(out, "%s",
271 				    archive_entry_pathname(entry));
272 			else
273 				list_item_verbose(bsdtar, out, entry);
274 			fflush(out);
275 			r = archive_read_data_skip(a);
276 			if (r == ARCHIVE_WARN) {
277 				fprintf(out, "\n");
278 				lafe_warnc(0, "%s",
279 				    archive_error_string(a));
280 			}
281 			if (r == ARCHIVE_RETRY) {
282 				fprintf(out, "\n");
283 				lafe_warnc(0, "%s",
284 				    archive_error_string(a));
285 			}
286 			if (r == ARCHIVE_FATAL) {
287 				fprintf(out, "\n");
288 				lafe_warnc(0, "%s",
289 				    archive_error_string(a));
290 				bsdtar->return_value = 1;
291 				break;
292 			}
293 			fprintf(out, "\n");
294 		} else {
295 			/* Note: some rewrite failures prevent extraction. */
296 			if (edit_pathname(bsdtar, entry))
297 				continue; /* Excluded by a rewrite failure. */
298 
299 			if (bsdtar->option_interactive &&
300 			    !yes("extract '%s'", archive_entry_pathname(entry)))
301 				continue;
302 
303 			/*
304 			 * Format here is from SUSv2, including the
305 			 * deferred '\n'.
306 			 */
307 			if (bsdtar->verbose) {
308 				safe_fprintf(stderr, "x %s",
309 				    archive_entry_pathname(entry));
310 				fflush(stderr);
311 			}
312 
313 			// TODO siginfo_printinfo(bsdtar, 0);
314 
315 			if (bsdtar->option_stdout)
316 				r = archive_read_data_into_fd(a, 1);
317 			else
318 				r = archive_read_extract(a, entry,
319 				    bsdtar->extract_flags);
320 			if (r != ARCHIVE_OK) {
321 				if (!bsdtar->verbose)
322 					safe_fprintf(stderr, "%s",
323 					    archive_entry_pathname(entry));
324 				safe_fprintf(stderr, ": %s",
325 				    archive_error_string(a));
326 				if (!bsdtar->verbose)
327 					fprintf(stderr, "\n");
328 				bsdtar->return_value = 1;
329 			}
330 			if (bsdtar->verbose)
331 				fprintf(stderr, "\n");
332 			if (r == ARCHIVE_FATAL)
333 				break;
334 		}
335 	}
336 
337 
338 	r = archive_read_close(a);
339 	if (r != ARCHIVE_OK)
340 		lafe_warnc(0, "%s", archive_error_string(a));
341 	if (r <= ARCHIVE_WARN)
342 		bsdtar->return_value = 1;
343 
344 	if (bsdtar->verbose > 2)
345 		fprintf(stdout, "Archive Format: %s,  Compression: %s\n",
346 		    archive_format_name(a), archive_compression_name(a));
347 
348 	archive_read_finish(a);
349 }
350 
351 
352 /*
353  * Display information about the current file.
354  *
355  * The format here roughly duplicates the output of 'ls -l'.
356  * This is based on SUSv2, where 'tar tv' is documented as
357  * listing additional information in an "unspecified format,"
358  * and 'pax -l' is documented as using the same format as 'ls -l'.
359  */
360 static void
361 list_item_verbose(struct bsdtar *bsdtar, FILE *out, struct archive_entry *entry)
362 {
363 	char			 tmp[100];
364 	size_t			 w;
365 	const char		*p;
366 	const char		*fmt;
367 	time_t			 tim;
368 	static time_t		 now;
369 
370 	/*
371 	 * We avoid collecting the entire list in memory at once by
372 	 * listing things as we see them.  However, that also means we can't
373 	 * just pre-compute the field widths.  Instead, we start with guesses
374 	 * and just widen them as necessary.  These numbers are completely
375 	 * arbitrary.
376 	 */
377 	if (!bsdtar->u_width) {
378 		bsdtar->u_width = 6;
379 		bsdtar->gs_width = 13;
380 	}
381 	if (!now)
382 		time(&now);
383 	fprintf(out, "%s %d ",
384 	    archive_entry_strmode(entry),
385 	    archive_entry_nlink(entry));
386 
387 	/* Use uname if it's present, else uid. */
388 	p = archive_entry_uname(entry);
389 	if ((p == NULL) || (*p == '\0')) {
390 		sprintf(tmp, "%lu ",
391 		    (unsigned long)archive_entry_uid(entry));
392 		p = tmp;
393 	}
394 	w = strlen(p);
395 	if (w > bsdtar->u_width)
396 		bsdtar->u_width = w;
397 	fprintf(out, "%-*s ", (int)bsdtar->u_width, p);
398 
399 	/* Use gname if it's present, else gid. */
400 	p = archive_entry_gname(entry);
401 	if (p != NULL && p[0] != '\0') {
402 		fprintf(out, "%s", p);
403 		w = strlen(p);
404 	} else {
405 		sprintf(tmp, "%lu",
406 		    (unsigned long)archive_entry_gid(entry));
407 		w = strlen(tmp);
408 		fprintf(out, "%s", tmp);
409 	}
410 
411 	/*
412 	 * Print device number or file size, right-aligned so as to make
413 	 * total width of group and devnum/filesize fields be gs_width.
414 	 * If gs_width is too small, grow it.
415 	 */
416 	if (archive_entry_filetype(entry) == AE_IFCHR
417 	    || archive_entry_filetype(entry) == AE_IFBLK) {
418 		sprintf(tmp, "%lu,%lu",
419 		    (unsigned long)archive_entry_rdevmajor(entry),
420 		    (unsigned long)archive_entry_rdevminor(entry));
421 	} else {
422 		strcpy(tmp, tar_i64toa(archive_entry_size(entry)));
423 	}
424 	if (w + strlen(tmp) >= bsdtar->gs_width)
425 		bsdtar->gs_width = w+strlen(tmp)+1;
426 	fprintf(out, "%*s", (int)(bsdtar->gs_width - w), tmp);
427 
428 	/* Format the time using 'ls -l' conventions. */
429 	tim = archive_entry_mtime(entry);
430 #define HALF_YEAR (time_t)365 * 86400 / 2
431 #if defined(_WIN32) && !defined(__CYGWIN__)
432 #define DAY_FMT  "%d"  /* Windows' strftime function does not support %e format. */
433 #else
434 #define DAY_FMT  "%e"  /* Day number without leading zeros */
435 #endif
436 	if (tim < now - HALF_YEAR || tim > now + HALF_YEAR)
437 		fmt = bsdtar->day_first ? DAY_FMT " %b  %Y" : "%b " DAY_FMT "  %Y";
438 	else
439 		fmt = bsdtar->day_first ? DAY_FMT " %b %H:%M" : "%b " DAY_FMT " %H:%M";
440 	strftime(tmp, sizeof(tmp), fmt, localtime(&tim));
441 	fprintf(out, " %s ", tmp);
442 	safe_fprintf(out, "%s", archive_entry_pathname(entry));
443 
444 	/* Extra information for links. */
445 	if (archive_entry_hardlink(entry)) /* Hard link */
446 		safe_fprintf(out, " link to %s",
447 		    archive_entry_hardlink(entry));
448 	else if (archive_entry_symlink(entry)) /* Symbolic link */
449 		safe_fprintf(out, " -> %s", archive_entry_symlink(entry));
450 }
451