xref: /freebsd/contrib/libarchive/tar/cmdline.c (revision e7437ae907c89bf85a99c5cbb7ddd194a1ff1354)
1 /*-
2  * Copyright (c) 2003-2008 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 /*
27  * Command line parser for tar.
28  */
29 
30 #include "bsdtar_platform.h"
31 __FBSDID("$FreeBSD$");
32 
33 #ifdef HAVE_ERRNO_H
34 #include <errno.h>
35 #endif
36 #ifdef HAVE_STDLIB_H
37 #include <stdlib.h>
38 #endif
39 #ifdef HAVE_STRING_H
40 #include <string.h>
41 #endif
42 
43 #include "bsdtar.h"
44 #include "err.h"
45 
46 /*
47  * Short options for tar.  Please keep this sorted.
48  */
49 static const char *short_options
50 	= "aBb:C:cf:HhI:JjkLlmnOoPpqrSs:T:tUuvW:wX:xyZz";
51 
52 /*
53  * Long options for tar.  Please keep this list sorted.
54  *
55  * The symbolic names for options that lack a short equivalent are
56  * defined in bsdtar.h.  Also note that so far I've found no need
57  * to support optional arguments to long options.  That would be
58  * a small change to the code below.
59  */
60 
61 static const struct bsdtar_option {
62 	const char *name;
63 	int required;      /* 1 if this option requires an argument. */
64 	int equivalent;    /* Equivalent short option. */
65 } tar_longopts[] = {
66 	{ "absolute-paths",       0, 'P' },
67 	{ "append",               0, 'r' },
68 	{ "acls",                 0, OPTION_ACLS },
69 	{ "auto-compress",        0, 'a' },
70 	{ "b64encode",            0, OPTION_B64ENCODE },
71 	{ "block-size",           1, 'b' },
72 	{ "blocking-factor",	  1, 'b' },
73 	{ "bunzip2",              0, 'j' },
74 	{ "bzip",                 0, 'j' },
75 	{ "bzip2",                0, 'j' },
76 	{ "cd",                   1, 'C' },
77 	{ "check-links",          0, OPTION_CHECK_LINKS },
78 	{ "chroot",               0, OPTION_CHROOT },
79 	{ "clear-nochange-fflags", 0, OPTION_CLEAR_NOCHANGE_FFLAGS },
80 	{ "compress",             0, 'Z' },
81 	{ "confirmation",         0, 'w' },
82 	{ "create",               0, 'c' },
83 	{ "dereference",	  0, 'L' },
84 	{ "directory",            1, 'C' },
85 	{ "disable-copyfile",	  0, OPTION_NO_MAC_METADATA },
86 	{ "exclude",              1, OPTION_EXCLUDE },
87 	{ "exclude-from",         1, 'X' },
88 	{ "exclude-vcs",	  0, OPTION_EXCLUDE_VCS },
89 	{ "extract",              0, 'x' },
90 	{ "fast-read",            0, 'q' },
91 	{ "fflags",               0, OPTION_FFLAGS },
92 	{ "file",                 1, 'f' },
93 	{ "files-from",           1, 'T' },
94 	{ "format",               1, OPTION_FORMAT },
95 	{ "gid",		  1, OPTION_GID },
96 	{ "gname",		  1, OPTION_GNAME },
97 	{ "grzip",                0, OPTION_GRZIP },
98 	{ "gunzip",               0, 'z' },
99 	{ "gzip",                 0, 'z' },
100 	{ "help",                 0, OPTION_HELP },
101 	{ "hfsCompression",       0, OPTION_HFS_COMPRESSION },
102 	{ "ignore-zeros",         0, OPTION_IGNORE_ZEROS },
103 	{ "include",              1, OPTION_INCLUDE },
104 	{ "insecure",             0, 'P' },
105 	{ "interactive",          0, 'w' },
106 	{ "keep-newer-files",     0, OPTION_KEEP_NEWER_FILES },
107 	{ "keep-old-files",       0, 'k' },
108 	{ "list",                 0, 't' },
109 	{ "lrzip",                0, OPTION_LRZIP },
110 	{ "lz4",                  0, OPTION_LZ4 },
111 	{ "lzip",                 0, OPTION_LZIP },
112 	{ "lzma",                 0, OPTION_LZMA },
113 	{ "lzop",                 0, OPTION_LZOP },
114 	{ "mac-metadata",         0, OPTION_MAC_METADATA },
115 	{ "modification-time",    0, 'm' },
116 	{ "newer",		  1, OPTION_NEWER_CTIME },
117 	{ "newer-ctime",	  1, OPTION_NEWER_CTIME },
118 	{ "newer-ctime-than",	  1, OPTION_NEWER_CTIME_THAN },
119 	{ "newer-mtime",	  1, OPTION_NEWER_MTIME },
120 	{ "newer-mtime-than",	  1, OPTION_NEWER_MTIME_THAN },
121 	{ "newer-than",		  1, OPTION_NEWER_CTIME_THAN },
122 	{ "no-acls",              0, OPTION_NO_ACLS },
123 	{ "no-fflags",            0, OPTION_NO_FFLAGS },
124 	{ "no-mac-metadata",      0, OPTION_NO_MAC_METADATA },
125 	{ "no-read-sparse",       0, OPTION_NO_READ_SPARSE },
126 	{ "no-recursion",         0, 'n' },
127 	{ "no-safe-writes",	  0, OPTION_NO_SAFE_WRITES },
128 	{ "no-same-owner",	  0, OPTION_NO_SAME_OWNER },
129 	{ "no-same-permissions",  0, OPTION_NO_SAME_PERMISSIONS },
130 	{ "no-xattr",             0, OPTION_NO_XATTRS },
131 	{ "no-xattrs",            0, OPTION_NO_XATTRS },
132 	{ "nodump",               0, OPTION_NODUMP },
133 	{ "nopreserveHFSCompression",0, OPTION_NOPRESERVE_HFS_COMPRESSION },
134 	{ "norecurse",            0, 'n' },
135 	{ "null",		  0, OPTION_NULL },
136 	{ "numeric-owner",	  0, OPTION_NUMERIC_OWNER },
137 	{ "older",		  1, OPTION_OLDER_CTIME },
138 	{ "older-ctime",	  1, OPTION_OLDER_CTIME },
139 	{ "older-ctime-than",	  1, OPTION_OLDER_CTIME_THAN },
140 	{ "older-mtime",	  1, OPTION_OLDER_MTIME },
141 	{ "older-mtime-than",	  1, OPTION_OLDER_MTIME_THAN },
142 	{ "older-than",		  1, OPTION_OLDER_CTIME_THAN },
143 	{ "one-file-system",	  0, OPTION_ONE_FILE_SYSTEM },
144 	{ "options",              1, OPTION_OPTIONS },
145 	{ "passphrase",		  1, OPTION_PASSPHRASE },
146 	{ "posix",		  0, OPTION_POSIX },
147 	{ "preserve-permissions", 0, 'p' },
148 	{ "read-full-blocks",	  0, 'B' },
149 	{ "read-sparse",	  0, OPTION_READ_SPARSE },
150 	{ "safe-writes",	  0, OPTION_SAFE_WRITES },
151 	{ "same-owner",	          0, OPTION_SAME_OWNER },
152 	{ "same-permissions",     0, 'p' },
153 	{ "strip-components",	  1, OPTION_STRIP_COMPONENTS },
154 	{ "to-stdout",            0, 'O' },
155 	{ "totals",		  0, OPTION_TOTALS },
156 	{ "uid",		  1, OPTION_UID },
157 	{ "uname",		  1, OPTION_UNAME },
158 	{ "uncompress",           0, 'Z' },
159 	{ "unlink",		  0, 'U' },
160 	{ "unlink-first",	  0, 'U' },
161 	{ "update",               0, 'u' },
162 	{ "use-compress-program", 1, OPTION_USE_COMPRESS_PROGRAM },
163 	{ "uuencode",             0, OPTION_UUENCODE },
164 	{ "verbose",              0, 'v' },
165 	{ "version",              0, OPTION_VERSION },
166 	{ "xattrs",               0, OPTION_XATTRS },
167 	{ "xz",                   0, 'J' },
168 	{ "zstd",                 0, OPTION_ZSTD },
169 	{ NULL, 0, 0 }
170 };
171 
172 /*
173  * This getopt implementation has two key features that common
174  * getopt_long() implementations lack.  Apart from those, it's a
175  * straightforward option parser, considerably simplified by not
176  * needing to support the wealth of exotic getopt_long() features.  It
177  * has, of course, been shamelessly tailored for bsdtar.  (If you're
178  * looking for a generic getopt_long() implementation for your
179  * project, I recommend Gregory Pietsch's public domain getopt_long()
180  * implementation.)  The two additional features are:
181  *
182  * Old-style tar arguments: The original tar implementation treated
183  * the first argument word as a list of single-character option
184  * letters.  All arguments follow as separate words.  For example,
185  *    tar xbf 32 /dev/tape
186  * Here, the "xbf" is three option letters, "32" is the argument for
187  * "b" and "/dev/tape" is the argument for "f".  We support this usage
188  * if the first command-line argument does not begin with '-'.  We
189  * also allow regular short and long options to follow, e.g.,
190  *    tar xbf 32 /dev/tape -P --format=pax
191  *
192  * -W long options: There's an obscure GNU convention (only rarely
193  * supported even there) that allows "-W option=argument" as an
194  * alternative way to support long options.  This was supported in
195  * early bsdtar as a way to access long options on platforms that did
196  * not support getopt_long() and is preserved here for backwards
197  * compatibility.  (Of course, if I'd started with a custom
198  * command-line parser from the beginning, I would have had normal
199  * long option support on every platform so that hack wouldn't have
200  * been necessary.  Oh, well.  Some mistakes you just have to live
201  * with.)
202  *
203  * TODO: We should be able to use this to pull files and intermingled
204  * options (such as -C) from the command line in write mode.  That
205  * will require a little rethinking of the argument handling in
206  * bsdtar.c.
207  *
208  * TODO: If we want to support arbitrary command-line options from -T
209  * input (as GNU tar does), we may need to extend this to handle option
210  * words from sources other than argv/argc.  I'm not really sure if I
211  * like that feature of GNU tar, so it's certainly not a priority.
212  */
213 
214 int
215 bsdtar_getopt(struct bsdtar *bsdtar)
216 {
217 	enum { state_start = 0, state_old_tar, state_next_word,
218 	       state_short, state_long };
219 
220 	const struct bsdtar_option *popt, *match = NULL, *match2 = NULL;
221 	const char *p, *long_prefix = "--";
222 	size_t optlength;
223 	int opt = '?';
224 	int required = 0;
225 
226 	bsdtar->argument = NULL;
227 
228 	/* First time through, initialize everything. */
229 	if (bsdtar->getopt_state == state_start) {
230 		/* Skip program name. */
231 		++bsdtar->argv;
232 		--bsdtar->argc;
233 		if (*bsdtar->argv == NULL)
234 			return (-1);
235 		/* Decide between "new style" and "old style" arguments. */
236 		if (bsdtar->argv[0][0] == '-') {
237 			bsdtar->getopt_state = state_next_word;
238 		} else {
239 			bsdtar->getopt_state = state_old_tar;
240 			bsdtar->getopt_word = *bsdtar->argv++;
241 			--bsdtar->argc;
242 		}
243 	}
244 
245 	/*
246 	 * We're parsing old-style tar arguments
247 	 */
248 	if (bsdtar->getopt_state == state_old_tar) {
249 		/* Get the next option character. */
250 		opt = *bsdtar->getopt_word++;
251 		if (opt == '\0') {
252 			/* New-style args can follow old-style. */
253 			bsdtar->getopt_state = state_next_word;
254 		} else {
255 			/* See if it takes an argument. */
256 			p = strchr(short_options, opt);
257 			if (p == NULL)
258 				return ('?');
259 			if (p[1] == ':') {
260 				bsdtar->argument = *bsdtar->argv;
261 				if (bsdtar->argument == NULL) {
262 					lafe_warnc(0,
263 					    "Option %c requires an argument",
264 					    opt);
265 					return ('?');
266 				}
267 				++bsdtar->argv;
268 				--bsdtar->argc;
269 			}
270 		}
271 	}
272 
273 	/*
274 	 * We're ready to look at the next word in argv.
275 	 */
276 	if (bsdtar->getopt_state == state_next_word) {
277 		/* No more arguments, so no more options. */
278 		if (bsdtar->argv[0] == NULL)
279 			return (-1);
280 		/* Doesn't start with '-', so no more options. */
281 		if (bsdtar->argv[0][0] != '-')
282 			return (-1);
283 		/* "--" marks end of options; consume it and return. */
284 		if (strcmp(bsdtar->argv[0], "--") == 0) {
285 			++bsdtar->argv;
286 			--bsdtar->argc;
287 			return (-1);
288 		}
289 		/* Get next word for parsing. */
290 		bsdtar->getopt_word = *bsdtar->argv++;
291 		--bsdtar->argc;
292 		if (bsdtar->getopt_word[1] == '-') {
293 			/* Set up long option parser. */
294 			bsdtar->getopt_state = state_long;
295 			bsdtar->getopt_word += 2; /* Skip leading '--' */
296 		} else {
297 			/* Set up short option parser. */
298 			bsdtar->getopt_state = state_short;
299 			++bsdtar->getopt_word;  /* Skip leading '-' */
300 		}
301 	}
302 
303 	/*
304 	 * We're parsing a group of POSIX-style single-character options.
305 	 */
306 	if (bsdtar->getopt_state == state_short) {
307 		/* Peel next option off of a group of short options. */
308 		opt = *bsdtar->getopt_word++;
309 		if (opt == '\0') {
310 			/* End of this group; recurse to get next option. */
311 			bsdtar->getopt_state = state_next_word;
312 			return bsdtar_getopt(bsdtar);
313 		}
314 
315 		/* Does this option take an argument? */
316 		p = strchr(short_options, opt);
317 		if (p == NULL)
318 			return ('?');
319 		if (p[1] == ':')
320 			required = 1;
321 
322 		/* If it takes an argument, parse that. */
323 		if (required) {
324 			/* If arg is run-in, bsdtar->getopt_word already points to it. */
325 			if (bsdtar->getopt_word[0] == '\0') {
326 				/* Otherwise, pick up the next word. */
327 				bsdtar->getopt_word = *bsdtar->argv;
328 				if (bsdtar->getopt_word == NULL) {
329 					lafe_warnc(0,
330 					    "Option -%c requires an argument",
331 					    opt);
332 					return ('?');
333 				}
334 				++bsdtar->argv;
335 				--bsdtar->argc;
336 			}
337 			if (opt == 'W') {
338 				bsdtar->getopt_state = state_long;
339 				long_prefix = "-W "; /* For clearer errors. */
340 			} else {
341 				bsdtar->getopt_state = state_next_word;
342 				bsdtar->argument = bsdtar->getopt_word;
343 			}
344 		}
345 	}
346 
347 	/* We're reading a long option, including -W long=arg convention. */
348 	if (bsdtar->getopt_state == state_long) {
349 		/* After this long option, we'll be starting a new word. */
350 		bsdtar->getopt_state = state_next_word;
351 
352 		/* Option name ends at '=' if there is one. */
353 		p = strchr(bsdtar->getopt_word, '=');
354 		if (p != NULL) {
355 			optlength = (size_t)(p - bsdtar->getopt_word);
356 			bsdtar->argument = (char *)(uintptr_t)(p + 1);
357 		} else {
358 			optlength = strlen(bsdtar->getopt_word);
359 		}
360 
361 		/* Search the table for an unambiguous match. */
362 		for (popt = tar_longopts; popt->name != NULL; popt++) {
363 			/* Short-circuit if first chars don't match. */
364 			if (popt->name[0] != bsdtar->getopt_word[0])
365 				continue;
366 			/* If option is a prefix of name in table, record it.*/
367 			if (strncmp(bsdtar->getopt_word, popt->name, optlength) == 0) {
368 				match2 = match; /* Record up to two matches. */
369 				match = popt;
370 				/* If it's an exact match, we're done. */
371 				if (strlen(popt->name) == optlength) {
372 					match2 = NULL; /* Forget the others. */
373 					break;
374 				}
375 			}
376 		}
377 
378 		/* Fail if there wasn't a unique match. */
379 		if (match == NULL) {
380 			lafe_warnc(0,
381 			    "Option %s%s is not supported",
382 			    long_prefix, bsdtar->getopt_word);
383 			return ('?');
384 		}
385 		if (match2 != NULL) {
386 			lafe_warnc(0,
387 			    "Ambiguous option %s%s (matches --%s and --%s)",
388 			    long_prefix, bsdtar->getopt_word, match->name, match2->name);
389 			return ('?');
390 		}
391 
392 		/* We've found a unique match; does it need an argument? */
393 		if (match->required) {
394 			/* Argument required: get next word if necessary. */
395 			if (bsdtar->argument == NULL) {
396 				bsdtar->argument = *bsdtar->argv;
397 				if (bsdtar->argument == NULL) {
398 					lafe_warnc(0,
399 					    "Option %s%s requires an argument",
400 					    long_prefix, match->name);
401 					return ('?');
402 				}
403 				++bsdtar->argv;
404 				--bsdtar->argc;
405 			}
406 		} else {
407 			/* Argument forbidden: fail if there is one. */
408 			if (bsdtar->argument != NULL) {
409 				lafe_warnc(0,
410 				    "Option %s%s does not allow an argument",
411 				    long_prefix, match->name);
412 				return ('?');
413 			}
414 		}
415 		return (match->equivalent);
416 	}
417 
418 	return (opt);
419 }
420