xref: /minix/external/bsd/file/dist/src/file.c (revision 0a6a1f1d)
1 /*	$NetBSD: file.c,v 1.7 2015/01/02 21:15:32 christos Exp $	*/
2 
3 /*
4  * Copyright (c) Ian F. Darwin 1986-1995.
5  * Software written by Ian F. Darwin and others;
6  * maintained 1995-present by Christos Zoulas and others.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice immediately at the beginning of the file, without modification,
13  *    this list of conditions, and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  *
18  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21  * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
22  * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28  * SUCH DAMAGE.
29  */
30 /*
31  * file - find type of a file or files - main program.
32  */
33 
34 #include "file.h"
35 
36 #ifndef	lint
37 #if 0
38 FILE_RCSID("@(#)$File: file.c,v 1.160 2014/12/16 23:18:40 christos Exp $")
39 #else
40 __RCSID("$NetBSD: file.c,v 1.7 2015/01/02 21:15:32 christos Exp $");
41 #endif
42 #endif	/* lint */
43 
44 #include "magic.h"
45 
46 #include <stdlib.h>
47 #include <unistd.h>
48 #include <string.h>
49 #ifdef RESTORE_TIME
50 # if (__COHERENT__ >= 0x420)
51 #  include <sys/utime.h>
52 # else
53 #  ifdef USE_UTIMES
54 #   include <sys/time.h>
55 #  else
56 #   include <utime.h>
57 #  endif
58 # endif
59 #endif
60 #ifdef HAVE_UNISTD_H
61 #include <unistd.h>	/* for read() */
62 #endif
63 #ifdef HAVE_WCHAR_H
64 #include <wchar.h>
65 #endif
66 
67 #if defined(HAVE_GETOPT_H) && defined(HAVE_STRUCT_OPTION)
68 #include <getopt.h>
69 #ifndef HAVE_GETOPT_LONG
70 int getopt_long(int argc, char * const *argv, const char *optstring, const struct option *longopts, int *longindex);
71 #endif
72 #else
73 #include "mygetopt.h"
74 #endif
75 
76 #ifdef S_IFLNK
77 #define FILE_FLAGS "-bcEhikLlNnprsvz0"
78 #else
79 #define FILE_FLAGS "-bcEiklNnprsvz0"
80 #endif
81 
82 # define USAGE  \
83     "Usage: %s [" FILE_FLAGS \
84 	"] [--apple] [--mime-encoding] [--mime-type]\n" \
85     "            [-e testname] [-F separator] [-f namefile] [-m magicfiles] " \
86     "file ...\n" \
87     "       %s -C [-m magicfiles]\n" \
88     "       %s [--help]\n"
89 
90 private int 		/* Global command-line options 		*/
91 	bflag = 0,	/* brief output format	 		*/
92 	nopad = 0,	/* Don't pad output			*/
93 	nobuffer = 0,   /* Do not buffer stdout 		*/
94 	nulsep = 0;	/* Append '\0' to the separator		*/
95 
96 private const char *separator = ":";	/* Default field separator	*/
97 private const struct option long_options[] = {
98 #define OPT(shortname, longname, opt, doc)      \
99     {longname, opt, NULL, shortname},
100 #define OPT_LONGONLY(longname, opt, doc)        \
101     {longname, opt, NULL, 0},
102 #include "file_opts.h"
103 #undef OPT
104 #undef OPT_LONGONLY
105     {0, 0, NULL, 0}
106 };
107 #define OPTSTRING	"bcCde:Ef:F:hiklLm:nNpP:rsvz0"
108 
109 private const struct {
110 	const char *name;
111 	int value;
112 } nv[] = {
113 	{ "apptype",	MAGIC_NO_CHECK_APPTYPE },
114 	{ "ascii",	MAGIC_NO_CHECK_ASCII },
115 	{ "cdf",	MAGIC_NO_CHECK_CDF },
116 	{ "compress",	MAGIC_NO_CHECK_COMPRESS },
117 	{ "elf",	MAGIC_NO_CHECK_ELF },
118 	{ "encoding",	MAGIC_NO_CHECK_ENCODING },
119 	{ "soft",	MAGIC_NO_CHECK_SOFT },
120 	{ "tar",	MAGIC_NO_CHECK_TAR },
121 	{ "text",	MAGIC_NO_CHECK_TEXT },	/* synonym for ascii */
122 	{ "tokens",	MAGIC_NO_CHECK_TOKENS }, /* OBSOLETE: ignored for backwards compatibility */
123 };
124 
125 private struct {
126 	const char *name;
127 	int tag;
128 	size_t value;
129 } pm[] = {
130 	{ "indir",	MAGIC_PARAM_INDIR_MAX, 0 },
131 	{ "name",	MAGIC_PARAM_NAME_MAX, 0 },
132 	{ "elf_phnum",	MAGIC_PARAM_ELF_PHNUM_MAX, 0 },
133 	{ "elf_shnum",	MAGIC_PARAM_ELF_SHNUM_MAX, 0 },
134 	{ "elf_notes",	MAGIC_PARAM_ELF_NOTES_MAX, 0 },
135 };
136 
137 private char *progname;		/* used throughout 		*/
138 
139 #ifdef __dead
140 __dead
141 #endif
142 private void usage(void);
143 private void docprint(const char *);
144 #ifdef __dead
145 __dead
146 #endif
147 private void help(void);
148 
149 private int unwrap(struct magic_set *, const char *);
150 private int process(struct magic_set *ms, const char *, int);
151 private struct magic_set *load(const char *, int);
152 private void setparam(const char *);
153 private void applyparam(magic_t);
154 
155 
156 /*
157  * main - parse arguments and handle options
158  */
159 int
main(int argc,char * argv[])160 main(int argc, char *argv[])
161 {
162 	int c;
163 	size_t i;
164 	int action = 0, didsomefiles = 0, errflg = 0;
165 	int flags = 0, e = 0;
166 	struct magic_set *magic = NULL;
167 	int longindex;
168 	const char *magicfile = NULL;		/* where the magic is	*/
169 
170 	/* makes islower etc work for other langs */
171 #ifdef HAVE_SETLOCALE
172 	(void)setlocale(LC_CTYPE, "");
173 #endif
174 
175 #ifdef __EMX__
176 	/* sh-like wildcard expansion! Shouldn't hurt at least ... */
177 	_wildcard(&argc, &argv);
178 #endif
179 
180 	if ((progname = strrchr(argv[0], '/')) != NULL)
181 		progname++;
182 	else
183 		progname = argv[0];
184 
185 #ifdef S_IFLNK
186 	flags |= getenv("POSIXLY_CORRECT") ? MAGIC_SYMLINK : 0;
187 #endif
188 	while ((c = getopt_long(argc, argv, OPTSTRING, long_options,
189 	    &longindex)) != -1)
190 		switch (c) {
191 		case 0 :
192 			switch (longindex) {
193 			case 0:
194 				help();
195 				break;
196 			case 10:
197 				flags |= MAGIC_APPLE;
198 				break;
199 			case 11:
200 				flags |= MAGIC_MIME_TYPE;
201 				break;
202 			case 12:
203 				flags |= MAGIC_MIME_ENCODING;
204 				break;
205 			}
206 			break;
207 		case '0':
208 			nulsep = 1;
209 			break;
210 		case 'b':
211 			bflag++;
212 			break;
213 		case 'c':
214 			action = FILE_CHECK;
215 			break;
216 		case 'C':
217 			action = FILE_COMPILE;
218 			break;
219 		case 'd':
220 			flags |= MAGIC_DEBUG|MAGIC_CHECK;
221 			break;
222 		case 'E':
223 			flags |= MAGIC_ERROR;
224 			break;
225 		case 'e':
226 			for (i = 0; i < sizeof(nv) / sizeof(nv[0]); i++)
227 				if (strcmp(nv[i].name, optarg) == 0)
228 					break;
229 
230 			if (i == sizeof(nv) / sizeof(nv[0]))
231 				errflg++;
232 			else
233 				flags |= nv[i].value;
234 			break;
235 
236 		case 'f':
237 			if(action)
238 				usage();
239 			if (magic == NULL)
240 				if ((magic = load(magicfile, flags)) == NULL)
241 					return 1;
242 			e |= unwrap(magic, optarg);
243 			++didsomefiles;
244 			break;
245 		case 'F':
246 			separator = optarg;
247 			break;
248 		case 'i':
249 			flags |= MAGIC_MIME;
250 			break;
251 		case 'k':
252 			flags |= MAGIC_CONTINUE;
253 			break;
254 		case 'l':
255 			action = FILE_LIST;
256 			break;
257 		case 'm':
258 			magicfile = optarg;
259 			break;
260 		case 'n':
261 			++nobuffer;
262 			break;
263 		case 'N':
264 			++nopad;
265 			break;
266 #if defined(HAVE_UTIME) || defined(HAVE_UTIMES)
267 		case 'p':
268 			flags |= MAGIC_PRESERVE_ATIME;
269 			break;
270 #endif
271 		case 'P':
272 			setparam(optarg);
273 			break;
274 		case 'r':
275 			flags |= MAGIC_RAW;
276 			break;
277 			break;
278 		case 's':
279 			flags |= MAGIC_DEVICES;
280 			break;
281 		case 'v':
282 			if (magicfile == NULL)
283 				magicfile = magic_getpath(magicfile, action);
284 			(void)fprintf(stdout, "%s-%s\n", progname, VERSION);
285 			(void)fprintf(stdout, "magic file from %s\n",
286 				       magicfile);
287 			return 0;
288 		case 'z':
289 			flags |= MAGIC_COMPRESS;
290 			break;
291 #ifdef S_IFLNK
292 		case 'L':
293 			flags |= MAGIC_SYMLINK;
294 			break;
295 		case 'h':
296 			flags &= ~MAGIC_SYMLINK;
297 			break;
298 #endif
299 		case '?':
300 		default:
301 			errflg++;
302 			break;
303 		}
304 
305 	if (errflg) {
306 		usage();
307 	}
308 	if (e)
309 		return e;
310 
311 	if (MAGIC_VERSION != magic_version())
312 		(void)fprintf(stderr, "%s: compiled magic version [%d] "
313 		    "does not match with shared library magic version [%d]\n",
314 		    progname, MAGIC_VERSION, magic_version());
315 
316 	switch(action) {
317 	case FILE_CHECK:
318 	case FILE_COMPILE:
319 	case FILE_LIST:
320 		/*
321 		 * Don't try to check/compile ~/.magic unless we explicitly
322 		 * ask for it.
323 		 */
324 		magic = magic_open(flags|MAGIC_CHECK);
325 		if (magic == NULL) {
326 			(void)fprintf(stderr, "%s: %s\n", progname,
327 			    strerror(errno));
328 			return 1;
329 		}
330 
331 
332 		switch(action) {
333 		case FILE_CHECK:
334 			c = magic_check(magic, magicfile);
335 			break;
336 		case FILE_COMPILE:
337 			c = magic_compile(magic, magicfile);
338 			break;
339 		case FILE_LIST:
340 			c = magic_list(magic, magicfile);
341 			break;
342 		default:
343 			abort();
344 		}
345 		if (c == -1) {
346 			(void)fprintf(stderr, "%s: %s\n", progname,
347 			    magic_error(magic));
348 			return 1;
349 		}
350 		return 0;
351 	default:
352 		if (magic == NULL)
353 			if ((magic = load(magicfile, flags)) == NULL)
354 				return 1;
355 		applyparam(magic);
356 	}
357 
358 	if (optind == argc) {
359 		if (!didsomefiles)
360 			usage();
361 	}
362 	else {
363 		size_t j, wid, nw;
364 		for (wid = 0, j = (size_t)optind; j < (size_t)argc; j++) {
365 			nw = file_mbswidth(argv[j]);
366 			if (nw > wid)
367 				wid = nw;
368 		}
369 		/*
370 		 * If bflag is only set twice, set it depending on
371 		 * number of files [this is undocumented, and subject to change]
372 		 */
373 		if (bflag == 2) {
374 			bflag = optind >= argc - 1;
375 		}
376 		for (; optind < argc; optind++)
377 			e |= process(magic, argv[optind], wid);
378 	}
379 
380 	if (magic)
381 		magic_close(magic);
382 	return e;
383 }
384 
385 private void
applyparam(magic_t magic)386 applyparam(magic_t magic)
387 {
388 	size_t i;
389 
390 	for (i = 0; i < __arraycount(pm); i++) {
391 		if (pm[i].value == 0)
392 			continue;
393 		if (magic_setparam(magic, pm[i].tag, &pm[i].value) == -1) {
394 			(void)fprintf(stderr, "%s: Can't set %s %s\n", progname,
395 				pm[i].name, strerror(errno));
396 			exit(1);
397 		}
398 	}
399 }
400 
401 private void
setparam(const char * p)402 setparam(const char *p)
403 {
404 	size_t i;
405 	char *s;
406 
407 	if ((s = strchr(p, '=')) == NULL)
408 		goto badparm;
409 
410 	for (i = 0; i < __arraycount(pm); i++) {
411 		if (strncmp(p, pm[i].name, s - p) != 0)
412 			continue;
413 		pm[i].value = atoi(s + 1);
414 		return;
415 	}
416 badparm:
417 	(void)fprintf(stderr, "%s: Unknown param %s\n", progname, p);
418 	exit(1);
419 }
420 
421 private struct magic_set *
422 /*ARGSUSED*/
load(const char * magicfile,int flags)423 load(const char *magicfile, int flags)
424 {
425 	struct magic_set *magic = magic_open(flags);
426 	if (magic == NULL) {
427 		(void)fprintf(stderr, "%s: %s\n", progname, strerror(errno));
428 		return NULL;
429 	}
430 	if (magic_load(magic, magicfile) == -1) {
431 		(void)fprintf(stderr, "%s: %s\n",
432 		    progname, magic_error(magic));
433 		magic_close(magic);
434 		return NULL;
435 	}
436 	return magic;
437 }
438 
439 /*
440  * unwrap -- read a file of filenames, do each one.
441  */
442 private int
unwrap(struct magic_set * ms,const char * fn)443 unwrap(struct magic_set *ms, const char *fn)
444 {
445 	FILE *f;
446 	ssize_t len;
447 	char *line = NULL;
448 	size_t llen = 0;
449 	int wid = 0, cwid;
450 	int e = 0;
451 
452 	if (strcmp("-", fn) == 0) {
453 		f = stdin;
454 		wid = 1;
455 	} else {
456 		if ((f = fopen(fn, "r")) == NULL) {
457 			(void)fprintf(stderr, "%s: Cannot open `%s' (%s).\n",
458 			    progname, fn, strerror(errno));
459 			return 1;
460 		}
461 
462 		while ((len = getline(&line, &llen, f)) > 0) {
463 			if (line[len - 1] == '\n')
464 				line[len - 1] = '\0';
465 			cwid = file_mbswidth(line);
466 			if (cwid > wid)
467 				wid = cwid;
468 		}
469 
470 		rewind(f);
471 	}
472 
473 	while ((len = getline(&line, &llen, f)) > 0) {
474 		if (line[len - 1] == '\n')
475 			line[len - 1] = '\0';
476 		e |= process(ms, line, wid);
477 		if(nobuffer)
478 			(void)fflush(stdout);
479 	}
480 
481 	free(line);
482 	(void)fclose(f);
483 	return e;
484 }
485 
486 /*
487  * Called for each input file on the command line (or in a list of files)
488  */
489 private int
process(struct magic_set * ms,const char * inname,int wid)490 process(struct magic_set *ms, const char *inname, int wid)
491 {
492 	const char *type;
493 	int std_in = strcmp(inname, "-") == 0;
494 
495 	if (wid > 0 && !bflag) {
496 		(void)printf("%s", std_in ? "/dev/stdin" : inname);
497 		if (nulsep)
498 			(void)putc('\0', stdout);
499 		(void)printf("%s", separator);
500 		(void)printf("%*s ",
501 		    (int) (nopad ? 0 : (wid - file_mbswidth(inname))), "");
502 	}
503 
504 	type = magic_file(ms, std_in ? NULL : inname);
505 	if (type == NULL) {
506 		(void)printf("ERROR: %s\n", magic_error(ms));
507 		return 1;
508 	} else {
509 		(void)printf("%s\n", type);
510 		return 0;
511 	}
512 }
513 
514 protected size_t
file_mbswidth(const char * s)515 file_mbswidth(const char *s)
516 {
517 #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
518 	size_t bytesconsumed, old_n, n, width = 0;
519 	mbstate_t state;
520 	wchar_t nextchar;
521 	(void)memset(&state, 0, sizeof(mbstate_t));
522 	old_n = n = strlen(s);
523 
524 	while (n > 0) {
525 		bytesconsumed = mbrtowc(&nextchar, s, n, &state);
526 		if (bytesconsumed == (size_t)(-1) ||
527 		    bytesconsumed == (size_t)(-2)) {
528 			/* Something went wrong, return something reasonable */
529 			return old_n;
530 		}
531 		if (s[0] == '\n') {
532 			/*
533 			 * do what strlen() would do, so that caller
534 			 * is always right
535 			 */
536 			width++;
537 		} else {
538 			int w = wcwidth(nextchar);
539 			if (w > 0)
540 				width += w;
541 		}
542 
543 		s += bytesconsumed, n -= bytesconsumed;
544 	}
545 	return width;
546 #else
547 	return strlen(s);
548 #endif
549 }
550 
551 private void
usage(void)552 usage(void)
553 {
554 	(void)fprintf(stderr, USAGE, progname, progname, progname);
555 	exit(1);
556 }
557 
558 private void
docprint(const char * opts)559 docprint(const char *opts)
560 {
561 	size_t i;
562 	int comma;
563 	char *sp, *p;
564 
565 	p = strstr(opts, "%o");
566 	if (p == NULL) {
567 		fprintf(stdout, "%s", opts);
568 		return;
569 	}
570 
571 	for (sp = p - 1; sp > opts && *sp == ' '; sp--)
572 		continue;
573 
574 	fprintf(stdout, "%.*s", (int)(p - opts), opts);
575 
576 	comma = 0;
577 	for (i = 0; i < __arraycount(nv); i++) {
578 		fprintf(stdout, "%s%s", comma++ ? ", " : "", nv[i].name);
579 		if (i && i % 5 == 0) {
580 			fprintf(stdout, ",\n%*s", (int)(p - sp - 1), "");
581 			comma = 0;
582 		}
583 	}
584 
585 	fprintf(stdout, "%s", opts + (p - opts) + 2);
586 }
587 
588 private void
help(void)589 help(void)
590 {
591 	(void)fputs(
592 "Usage: file [OPTION...] [FILE...]\n"
593 "Determine type of FILEs.\n"
594 "\n", stdout);
595 #define OPT(shortname, longname, opt, doc)      \
596 	fprintf(stdout, "  -%c, --" longname, shortname), \
597 	docprint(doc);
598 #define OPT_LONGONLY(longname, opt, doc)        \
599 	fprintf(stdout, "      --" longname),	\
600 	docprint(doc);
601 #include "file_opts.h"
602 #undef OPT
603 #undef OPT_LONGONLY
604 	fprintf(stdout, "\nReport bugs to http://bugs.gw.com/\n");
605 	exit(0);
606 }
607