xref: /freebsd/contrib/libarchive/tar/cmdline.c (revision 9bc300465e48e19d794d88d0c158a2adb92c7197)
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 
32 #ifdef HAVE_ERRNO_H
33 #include <errno.h>
34 #endif
35 #ifdef HAVE_STDLIB_H
36 #include <stdlib.h>
37 #endif
38 #ifdef HAVE_STRING_H
39 #include <string.h>
40 #endif
41 
42 #include "bsdtar.h"
43 #include "err.h"
44 
45 /*
46  * Short options for tar.  Please keep this sorted.
47  */
48 static const char *short_options
49 	= "aBb:C:cf:HhI:JjkLlmnOoPpqrSs:T:tUuvW:wX:xyZz";
50 
51 /*
52  * Long options for tar.  Please keep this list sorted.
53  *
54  * The symbolic names for options that lack a short equivalent are
55  * defined in bsdtar.h.  Also note that so far I've found no need
56  * to support optional arguments to long options.  That would be
57  * a small change to the code below.
58  */
59 
60 static const struct bsdtar_option {
61 	const char *name;
62 	int required;      /* 1 if this option requires an argument. */
63 	int equivalent;    /* Equivalent short option. */
64 } tar_longopts[] = {
65 	{ "absolute-paths",       0, 'P' },
66 	{ "append",               0, 'r' },
67 	{ "acls",                 0, OPTION_ACLS },
68 	{ "auto-compress",        0, 'a' },
69 	{ "b64encode",            0, OPTION_B64ENCODE },
70 	{ "block-size",           1, 'b' },
71 	{ "blocking-factor",	  1, 'b' },
72 	{ "bunzip2",              0, 'j' },
73 	{ "bzip",                 0, 'j' },
74 	{ "bzip2",                0, 'j' },
75 	{ "cd",                   1, 'C' },
76 	{ "check-links",          0, OPTION_CHECK_LINKS },
77 	{ "chroot",               0, OPTION_CHROOT },
78 	{ "clear-nochange-fflags", 0, OPTION_CLEAR_NOCHANGE_FFLAGS },
79 	{ "compress",             0, 'Z' },
80 	{ "confirmation",         0, 'w' },
81 	{ "create",               0, 'c' },
82 	{ "dereference",	  0, 'L' },
83 	{ "directory",            1, 'C' },
84 	{ "disable-copyfile",	  0, OPTION_NO_MAC_METADATA },
85 	{ "exclude",              1, OPTION_EXCLUDE },
86 	{ "exclude-from",         1, 'X' },
87 	{ "exclude-vcs",	  0, OPTION_EXCLUDE_VCS },
88 	{ "extract",              0, 'x' },
89 	{ "fast-read",            0, 'q' },
90 	{ "fflags",               0, OPTION_FFLAGS },
91 	{ "file",                 1, 'f' },
92 	{ "files-from",           1, 'T' },
93 	{ "format",               1, OPTION_FORMAT },
94 	{ "gid",		  1, OPTION_GID },
95 	{ "gname",		  1, OPTION_GNAME },
96 	{ "group",		  1, OPTION_GROUP },
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 	{ "owner",		  1, OPTION_OWNER },
146 	{ "passphrase",		  1, OPTION_PASSPHRASE },
147 	{ "posix",		  0, OPTION_POSIX },
148 	{ "preserve-permissions", 0, 'p' },
149 	{ "read-full-blocks",	  0, 'B' },
150 	{ "read-sparse",	  0, OPTION_READ_SPARSE },
151 	{ "safe-writes",	  0, OPTION_SAFE_WRITES },
152 	{ "same-owner",	          0, OPTION_SAME_OWNER },
153 	{ "same-permissions",     0, 'p' },
154 	{ "strip-components",	  1, OPTION_STRIP_COMPONENTS },
155 	{ "to-stdout",            0, 'O' },
156 	{ "totals",		  0, OPTION_TOTALS },
157 	{ "uid",		  1, OPTION_UID },
158 	{ "uname",		  1, OPTION_UNAME },
159 	{ "uncompress",           0, 'Z' },
160 	{ "unlink",		  0, 'U' },
161 	{ "unlink-first",	  0, 'U' },
162 	{ "update",               0, 'u' },
163 	{ "use-compress-program", 1, OPTION_USE_COMPRESS_PROGRAM },
164 	{ "uuencode",             0, OPTION_UUENCODE },
165 	{ "verbose",              0, 'v' },
166 	{ "version",              0, OPTION_VERSION },
167 	{ "xattrs",               0, OPTION_XATTRS },
168 	{ "xz",                   0, 'J' },
169 	{ "zstd",                 0, OPTION_ZSTD },
170 	{ NULL, 0, 0 }
171 };
172 
173 /*
174  * This getopt implementation has two key features that common
175  * getopt_long() implementations lack.  Apart from those, it's a
176  * straightforward option parser, considerably simplified by not
177  * needing to support the wealth of exotic getopt_long() features.  It
178  * has, of course, been shamelessly tailored for bsdtar.  (If you're
179  * looking for a generic getopt_long() implementation for your
180  * project, I recommend Gregory Pietsch's public domain getopt_long()
181  * implementation.)  The two additional features are:
182  *
183  * Old-style tar arguments: The original tar implementation treated
184  * the first argument word as a list of single-character option
185  * letters.  All arguments follow as separate words.  For example,
186  *    tar xbf 32 /dev/tape
187  * Here, the "xbf" is three option letters, "32" is the argument for
188  * "b" and "/dev/tape" is the argument for "f".  We support this usage
189  * if the first command-line argument does not begin with '-'.  We
190  * also allow regular short and long options to follow, e.g.,
191  *    tar xbf 32 /dev/tape -P --format=pax
192  *
193  * -W long options: There's an obscure GNU convention (only rarely
194  * supported even there) that allows "-W option=argument" as an
195  * alternative way to support long options.  This was supported in
196  * early bsdtar as a way to access long options on platforms that did
197  * not support getopt_long() and is preserved here for backwards
198  * compatibility.  (Of course, if I'd started with a custom
199  * command-line parser from the beginning, I would have had normal
200  * long option support on every platform so that hack wouldn't have
201  * been necessary.  Oh, well.  Some mistakes you just have to live
202  * with.)
203  *
204  * TODO: We should be able to use this to pull files and intermingled
205  * options (such as -C) from the command line in write mode.  That
206  * will require a little rethinking of the argument handling in
207  * bsdtar.c.
208  *
209  * TODO: If we want to support arbitrary command-line options from -T
210  * input (as GNU tar does), we may need to extend this to handle option
211  * words from sources other than argv/argc.  I'm not really sure if I
212  * like that feature of GNU tar, so it's certainly not a priority.
213  */
214 
215 int
216 bsdtar_getopt(struct bsdtar *bsdtar)
217 {
218 	enum { state_start = 0, state_old_tar, state_next_word,
219 	       state_short, state_long };
220 
221 	const struct bsdtar_option *popt, *match, *match2;
222 	const char *p, *long_prefix;
223 	size_t optlength;
224 	int opt;
225 	int required;
226 
227 again:
228 	match = NULL;
229 	match2 = NULL;
230 	long_prefix = "--";
231 	opt = '?';
232 	required = 0;
233 	bsdtar->argument = NULL;
234 
235 	/* First time through, initialize everything. */
236 	if (bsdtar->getopt_state == state_start) {
237 		/* Skip program name. */
238 		++bsdtar->argv;
239 		--bsdtar->argc;
240 		if (*bsdtar->argv == NULL)
241 			return (-1);
242 		/* Decide between "new style" and "old style" arguments. */
243 		if (bsdtar->argv[0][0] == '-') {
244 			bsdtar->getopt_state = state_next_word;
245 		} else {
246 			bsdtar->getopt_state = state_old_tar;
247 			bsdtar->getopt_word = *bsdtar->argv++;
248 			--bsdtar->argc;
249 		}
250 	}
251 
252 	/*
253 	 * We're parsing old-style tar arguments
254 	 */
255 	if (bsdtar->getopt_state == state_old_tar) {
256 		/* Get the next option character. */
257 		opt = *bsdtar->getopt_word++;
258 		if (opt == '\0') {
259 			/* New-style args can follow old-style. */
260 			bsdtar->getopt_state = state_next_word;
261 		} else {
262 			/* See if it takes an argument. */
263 			p = strchr(short_options, opt);
264 			if (p == NULL)
265 				return ('?');
266 			if (p[1] == ':') {
267 				bsdtar->argument = *bsdtar->argv;
268 				if (bsdtar->argument == NULL) {
269 					lafe_warnc(0,
270 					    "Option %c requires an argument",
271 					    opt);
272 					return ('?');
273 				}
274 				++bsdtar->argv;
275 				--bsdtar->argc;
276 			}
277 		}
278 	}
279 
280 	/*
281 	 * We're ready to look at the next word in argv.
282 	 */
283 	if (bsdtar->getopt_state == state_next_word) {
284 		/* No more arguments, so no more options. */
285 		if (bsdtar->argv[0] == NULL)
286 			return (-1);
287 		/* Doesn't start with '-', so no more options. */
288 		if (bsdtar->argv[0][0] != '-')
289 			return (-1);
290 		/* "--" marks end of options; consume it and return. */
291 		if (strcmp(bsdtar->argv[0], "--") == 0) {
292 			++bsdtar->argv;
293 			--bsdtar->argc;
294 			return (-1);
295 		}
296 		/* Get next word for parsing. */
297 		bsdtar->getopt_word = *bsdtar->argv++;
298 		--bsdtar->argc;
299 		if (bsdtar->getopt_word[1] == '-') {
300 			/* Set up long option parser. */
301 			bsdtar->getopt_state = state_long;
302 			bsdtar->getopt_word += 2; /* Skip leading '--' */
303 		} else {
304 			/* Set up short option parser. */
305 			bsdtar->getopt_state = state_short;
306 			++bsdtar->getopt_word;  /* Skip leading '-' */
307 		}
308 	}
309 
310 	/*
311 	 * We're parsing a group of POSIX-style single-character options.
312 	 */
313 	if (bsdtar->getopt_state == state_short) {
314 		/* Peel next option off of a group of short options. */
315 		opt = *bsdtar->getopt_word++;
316 		if (opt == '\0') {
317 			/* End of this group; recurse to get next option. */
318 			bsdtar->getopt_state = state_next_word;
319 			goto again;
320 		}
321 
322 		/* Does this option take an argument? */
323 		p = strchr(short_options, opt);
324 		if (p == NULL)
325 			return ('?');
326 		if (p[1] == ':')
327 			required = 1;
328 
329 		/* If it takes an argument, parse that. */
330 		if (required) {
331 			/* If arg is run-in, bsdtar->getopt_word already points to it. */
332 			if (bsdtar->getopt_word[0] == '\0') {
333 				/* Otherwise, pick up the next word. */
334 				bsdtar->getopt_word = *bsdtar->argv;
335 				if (bsdtar->getopt_word == NULL) {
336 					lafe_warnc(0,
337 					    "Option -%c requires an argument",
338 					    opt);
339 					return ('?');
340 				}
341 				++bsdtar->argv;
342 				--bsdtar->argc;
343 			}
344 			if (opt == 'W') {
345 				bsdtar->getopt_state = state_long;
346 				long_prefix = "-W "; /* For clearer errors. */
347 			} else {
348 				bsdtar->getopt_state = state_next_word;
349 				bsdtar->argument = bsdtar->getopt_word;
350 			}
351 		}
352 	}
353 
354 	/* We're reading a long option, including -W long=arg convention. */
355 	if (bsdtar->getopt_state == state_long) {
356 		/* After this long option, we'll be starting a new word. */
357 		bsdtar->getopt_state = state_next_word;
358 
359 		/* Option name ends at '=' if there is one. */
360 		p = strchr(bsdtar->getopt_word, '=');
361 		if (p != NULL) {
362 			optlength = (size_t)(p - bsdtar->getopt_word);
363 			bsdtar->argument = (char *)(uintptr_t)(p + 1);
364 		} else {
365 			optlength = strlen(bsdtar->getopt_word);
366 		}
367 
368 		/* Search the table for an unambiguous match. */
369 		for (popt = tar_longopts; popt->name != NULL; popt++) {
370 			/* Short-circuit if first chars don't match. */
371 			if (popt->name[0] != bsdtar->getopt_word[0])
372 				continue;
373 			/* If option is a prefix of name in table, record it.*/
374 			if (strncmp(bsdtar->getopt_word, popt->name, optlength) == 0) {
375 				match2 = match; /* Record up to two matches. */
376 				match = popt;
377 				/* If it's an exact match, we're done. */
378 				if (strlen(popt->name) == optlength) {
379 					match2 = NULL; /* Forget the others. */
380 					break;
381 				}
382 			}
383 		}
384 
385 		/* Fail if there wasn't a unique match. */
386 		if (match == NULL) {
387 			lafe_warnc(0,
388 			    "Option %s%s is not supported",
389 			    long_prefix, bsdtar->getopt_word);
390 			return ('?');
391 		}
392 		if (match2 != NULL) {
393 			lafe_warnc(0,
394 			    "Ambiguous option %s%s (matches --%s and --%s)",
395 			    long_prefix, bsdtar->getopt_word, match->name, match2->name);
396 			return ('?');
397 		}
398 
399 		/* We've found a unique match; does it need an argument? */
400 		if (match->required) {
401 			/* Argument required: get next word if necessary. */
402 			if (bsdtar->argument == NULL) {
403 				bsdtar->argument = *bsdtar->argv;
404 				if (bsdtar->argument == NULL) {
405 					lafe_warnc(0,
406 					    "Option %s%s requires an argument",
407 					    long_prefix, match->name);
408 					return ('?');
409 				}
410 				++bsdtar->argv;
411 				--bsdtar->argc;
412 			}
413 		} else {
414 			/* Argument forbidden: fail if there is one. */
415 			if (bsdtar->argument != NULL) {
416 				lafe_warnc(0,
417 				    "Option %s%s does not allow an argument",
418 				    long_prefix, match->name);
419 				return ('?');
420 			}
421 		}
422 		return (match->equivalent);
423 	}
424 
425 	return (opt);
426 }
427