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 "bsdcat_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 "bsdcat.h"
44#include "err.h"
45
46/*
47 * Short options for tar.  Please keep this sorted.
48 */
49static const char *short_options = "h";
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 bsdcat.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
60static const struct bsdcat_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	{ "help",                 0, 'h' },
66	{ "version",              0, OPTION_VERSION },
67	{ NULL, 0, 0 }
68};
69
70/*
71 * This getopt implementation has two key features that common
72 * getopt_long() implementations lack.  Apart from those, it's a
73 * straightforward option parser, considerably simplified by not
74 * needing to support the wealth of exotic getopt_long() features.  It
75 * has, of course, been shamelessly tailored for bsdcat.  (If you're
76 * looking for a generic getopt_long() implementation for your
77 * project, I recommend Gregory Pietsch's public domain getopt_long()
78 * implementation.)  The two additional features are:
79 *
80 * Old-style tar arguments: The original tar implementation treated
81 * the first argument word as a list of single-character option
82 * letters.  All arguments follow as separate words.  For example,
83 *    tar xbf 32 /dev/tape
84 * Here, the "xbf" is three option letters, "32" is the argument for
85 * "b" and "/dev/tape" is the argument for "f".  We support this usage
86 * if the first command-line argument does not begin with '-'.  We
87 * also allow regular short and long options to follow, e.g.,
88 *    tar xbf 32 /dev/tape -P --format=pax
89 *
90 * -W long options: There's an obscure GNU convention (only rarely
91 * supported even there) that allows "-W option=argument" as an
92 * alternative way to support long options.  This was supported in
93 * early bsdcat as a way to access long options on platforms that did
94 * not support getopt_long() and is preserved here for backwards
95 * compatibility.  (Of course, if I'd started with a custom
96 * command-line parser from the beginning, I would have had normal
97 * long option support on every platform so that hack wouldn't have
98 * been necessary.  Oh, well.  Some mistakes you just have to live
99 * with.)
100 *
101 * TODO: We should be able to use this to pull files and intermingled
102 * options (such as -C) from the command line in write mode.  That
103 * will require a little rethinking of the argument handling in
104 * bsdcat.c.
105 *
106 * TODO: If we want to support arbitrary command-line options from -T
107 * input (as GNU tar does), we may need to extend this to handle option
108 * words from sources other than argv/argc.  I'm not really sure if I
109 * like that feature of GNU tar, so it's certainly not a priority.
110 */
111
112int
113bsdcat_getopt(struct bsdcat *bsdcat)
114{
115	enum { state_start = 0, state_old_tar, state_next_word,
116	       state_short, state_long };
117
118	const struct bsdcat_option *popt, *match = NULL, *match2 = NULL;
119	const char *p, *long_prefix = "--";
120	size_t optlength;
121	int opt = '?';
122	int required = 0;
123
124	bsdcat->argument = NULL;
125
126	/* First time through, initialize everything. */
127	if (bsdcat->getopt_state == state_start) {
128		/* Skip program name. */
129		++bsdcat->argv;
130		--bsdcat->argc;
131		if (*bsdcat->argv == NULL)
132			return (-1);
133		/* Decide between "new style" and "old style" arguments. */
134		bsdcat->getopt_state = state_next_word;
135	}
136
137	/*
138	 * We're ready to look at the next word in argv.
139	 */
140	if (bsdcat->getopt_state == state_next_word) {
141		/* No more arguments, so no more options. */
142		if (bsdcat->argv[0] == NULL)
143			return (-1);
144		/* Doesn't start with '-', so no more options. */
145		if (bsdcat->argv[0][0] != '-')
146			return (-1);
147		/* "--" marks end of options; consume it and return. */
148		if (strcmp(bsdcat->argv[0], "--") == 0) {
149			++bsdcat->argv;
150			--bsdcat->argc;
151			return (-1);
152		}
153		/* Get next word for parsing. */
154		bsdcat->getopt_word = *bsdcat->argv++;
155		--bsdcat->argc;
156		if (bsdcat->getopt_word[1] == '-') {
157			/* Set up long option parser. */
158			bsdcat->getopt_state = state_long;
159			bsdcat->getopt_word += 2; /* Skip leading '--' */
160		} else {
161			/* Set up short option parser. */
162			bsdcat->getopt_state = state_short;
163			++bsdcat->getopt_word;  /* Skip leading '-' */
164		}
165	}
166
167	/*
168	 * We're parsing a group of POSIX-style single-character options.
169	 */
170	if (bsdcat->getopt_state == state_short) {
171		/* Peel next option off of a group of short options. */
172		opt = *bsdcat->getopt_word++;
173		if (opt == '\0') {
174			/* End of this group; recurse to get next option. */
175			bsdcat->getopt_state = state_next_word;
176			return bsdcat_getopt(bsdcat);
177		}
178
179		/* Does this option take an argument? */
180		p = strchr(short_options, opt);
181		if (p == NULL)
182			return ('?');
183		if (p[1] == ':')
184			required = 1;
185
186		/* If it takes an argument, parse that. */
187		if (required) {
188			/* If arg is run-in, bsdcat->getopt_word already points to it. */
189			if (bsdcat->getopt_word[0] == '\0') {
190				/* Otherwise, pick up the next word. */
191				bsdcat->getopt_word = *bsdcat->argv;
192				if (bsdcat->getopt_word == NULL) {
193					lafe_warnc(0,
194					    "Option -%c requires an argument",
195					    opt);
196					return ('?');
197				}
198				++bsdcat->argv;
199				--bsdcat->argc;
200			}
201			if (opt == 'W') {
202				bsdcat->getopt_state = state_long;
203				long_prefix = "-W "; /* For clearer errors. */
204			} else {
205				bsdcat->getopt_state = state_next_word;
206				bsdcat->argument = bsdcat->getopt_word;
207			}
208		}
209	}
210
211	/* We're reading a long option, including -W long=arg convention. */
212	if (bsdcat->getopt_state == state_long) {
213		/* After this long option, we'll be starting a new word. */
214		bsdcat->getopt_state = state_next_word;
215
216		/* Option name ends at '=' if there is one. */
217		p = strchr(bsdcat->getopt_word, '=');
218		if (p != NULL) {
219			optlength = (size_t)(p - bsdcat->getopt_word);
220			bsdcat->argument = (char *)(uintptr_t)(p + 1);
221		} else {
222			optlength = strlen(bsdcat->getopt_word);
223		}
224
225		/* Search the table for an unambiguous match. */
226		for (popt = tar_longopts; popt->name != NULL; popt++) {
227			/* Short-circuit if first chars don't match. */
228			if (popt->name[0] != bsdcat->getopt_word[0])
229				continue;
230			/* If option is a prefix of name in table, record it.*/
231			if (strncmp(bsdcat->getopt_word, popt->name, optlength) == 0) {
232				match2 = match; /* Record up to two matches. */
233				match = popt;
234				/* If it's an exact match, we're done. */
235				if (strlen(popt->name) == optlength) {
236					match2 = NULL; /* Forget the others. */
237					break;
238				}
239			}
240		}
241
242		/* Fail if there wasn't a unique match. */
243		if (match == NULL) {
244			lafe_warnc(0,
245			    "Option %s%s is not supported",
246			    long_prefix, bsdcat->getopt_word);
247			return ('?');
248		}
249		if (match2 != NULL) {
250			lafe_warnc(0,
251			    "Ambiguous option %s%s (matches --%s and --%s)",
252			    long_prefix, bsdcat->getopt_word, match->name, match2->name);
253			return ('?');
254		}
255
256		/* We've found a unique match; does it need an argument? */
257		if (match->required) {
258			/* Argument required: get next word if necessary. */
259			if (bsdcat->argument == NULL) {
260				bsdcat->argument = *bsdcat->argv;
261				if (bsdcat->argument == NULL) {
262					lafe_warnc(0,
263					    "Option %s%s requires an argument",
264					    long_prefix, match->name);
265					return ('?');
266				}
267				++bsdcat->argv;
268				--bsdcat->argc;
269			}
270		} else {
271			/* Argument forbidden: fail if there is one. */
272			if (bsdcat->argument != NULL) {
273				lafe_warnc(0,
274				    "Option %s%s does not allow an argument",
275				    long_prefix, match->name);
276				return ('?');
277			}
278		}
279		return (match->equivalent);
280	}
281
282	return (opt);
283}
284