1/*
2 * Copyright (c) Ian F. Darwin 1986-1995.
3 * Software written by Ian F. Darwin and others;
4 * maintained 1995-present by Christos Zoulas and others.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 *    notice immediately at the beginning of the file, without modification,
11 *    this list of conditions, and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 *    notice, this list of conditions and the following disclaimer in the
14 *    documentation and/or other materials provided with the distribution.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
20 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 */
28/*
29 * file - find type of a file or files - main program.
30 */
31
32#include "file.h"
33
34#ifndef	lint
35FILE_RCSID("@(#)$File: file.c,v 1.184 2019/08/03 11:51:59 christos Exp $")
36#endif	/* lint */
37
38#include "magic.h"
39
40#include <stdlib.h>
41#include <unistd.h>
42#include <string.h>
43#ifdef RESTORE_TIME
44# if (__COHERENT__ >= 0x420)
45#  include <sys/utime.h>
46# else
47#  ifdef USE_UTIMES
48#   include <sys/time.h>
49#  else
50#   include <utime.h>
51#  endif
52# endif
53#endif
54#ifdef HAVE_UNISTD_H
55#include <unistd.h>	/* for read() */
56#endif
57#ifdef HAVE_WCHAR_H
58#include <wchar.h>
59#endif
60
61#if defined(HAVE_GETOPT_H) && defined(HAVE_STRUCT_OPTION)
62# include <getopt.h>
63# ifndef HAVE_GETOPT_LONG
64int getopt_long(int, char * const *, const char *,
65    const struct option *, int *);
66# endif
67# else
68#  include "mygetopt.h"
69#endif
70
71#ifdef S_IFLNK
72# define IFLNK_h "h"
73# define IFLNK_L "L"
74#else
75# define IFLNK_h ""
76# define IFLNK_L ""
77#endif
78
79#define FILE_FLAGS	"bcCdE" IFLNK_h "ik" IFLNK_L "lNnprsSvzZ0"
80#define OPTSTRING	"bcCde:Ef:F:hiklLm:nNpP:rsSvzZ0"
81
82# define USAGE  \
83    "Usage: %s [-" FILE_FLAGS "] [--apple] [--extension] [--mime-encoding]\n" \
84    "            [--mime-type] [-e <testname>] [-F <separator>] " \
85    " [-f <namefile>]\n" \
86    "            [-m <magicfiles>] [-P <parameter=value>] <file> ...\n" \
87    "       %s -C [-m <magicfiles>]\n" \
88    "       %s [--help]\n"
89
90private 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
96private const char *separator = ":";	/* Default field separator	*/
97private const struct option long_options[] = {
98#define OPT_HELP		1
99#define OPT_APPLE		2
100#define OPT_EXTENSIONS		3
101#define OPT_MIME_TYPE		4
102#define OPT_MIME_ENCODING	5
103#define OPT(shortname, longname, opt, def, doc)      \
104    {longname, opt, NULL, shortname},
105#define OPT_LONGONLY(longname, opt, def, doc, id)        \
106    {longname, opt, NULL, id},
107#include "file_opts.h"
108#undef OPT
109#undef OPT_LONGONLY
110    {0, 0, NULL, 0}
111    };
112
113private const struct {
114	const char *name;
115	int value;
116} nv[] = {
117	{ "apptype",	MAGIC_NO_CHECK_APPTYPE },
118	{ "ascii",	MAGIC_NO_CHECK_ASCII },
119	{ "cdf",	MAGIC_NO_CHECK_CDF },
120	{ "compress",	MAGIC_NO_CHECK_COMPRESS },
121	{ "csv",	MAGIC_NO_CHECK_CSV },
122	{ "elf",	MAGIC_NO_CHECK_ELF },
123	{ "encoding",	MAGIC_NO_CHECK_ENCODING },
124	{ "soft",	MAGIC_NO_CHECK_SOFT },
125	{ "tar",	MAGIC_NO_CHECK_TAR },
126	{ "json",	MAGIC_NO_CHECK_JSON },
127	{ "text",	MAGIC_NO_CHECK_TEXT },	/* synonym for ascii */
128	{ "tokens",	MAGIC_NO_CHECK_TOKENS }, /* OBSOLETE: ignored for backwards compatibility */
129};
130
131private struct {
132	const char *name;
133	int tag;
134	size_t value;
135	int set;
136} pm[] = {
137	{ "indir",	MAGIC_PARAM_INDIR_MAX, 0, 0 },
138	{ "name",	MAGIC_PARAM_NAME_MAX, 0, 0 },
139	{ "elf_phnum",	MAGIC_PARAM_ELF_PHNUM_MAX, 0, 0 },
140	{ "elf_shnum",	MAGIC_PARAM_ELF_SHNUM_MAX, 0, 0 },
141	{ "elf_notes",	MAGIC_PARAM_ELF_NOTES_MAX, 0, 0 },
142	{ "regex",	MAGIC_PARAM_REGEX_MAX, 0, 0 },
143	{ "bytes",	MAGIC_PARAM_BYTES_MAX, 0, 0 },
144};
145
146private int posixly;
147
148#ifdef __dead
149__dead
150#endif
151private void usage(void);
152private void docprint(const char *, int);
153#ifdef __dead
154__dead
155#endif
156private void help(void);
157
158private int unwrap(struct magic_set *, const char *);
159private int process(struct magic_set *ms, const char *, int);
160private struct magic_set *load(const char *, int);
161private void setparam(const char *);
162private void applyparam(magic_t);
163
164
165/*
166 * main - parse arguments and handle options
167 */
168int
169main(int argc, char *argv[])
170{
171	int c;
172	size_t i;
173	int action = 0, didsomefiles = 0, errflg = 0;
174	int flags = 0, e = 0;
175#ifdef HAVE_LIBSECCOMP
176	int sandbox = 1;
177#endif
178	struct magic_set *magic = NULL;
179	int longindex;
180	const char *magicfile = NULL;		/* where the magic is	*/
181	char *progname;
182
183	/* makes islower etc work for other langs */
184	(void)setlocale(LC_CTYPE, "");
185
186#ifdef __EMX__
187	/* sh-like wildcard expansion! Shouldn't hurt at least ... */
188	_wildcard(&argc, &argv);
189#endif
190
191	if ((progname = strrchr(argv[0], '/')) != NULL)
192		progname++;
193	else
194		progname = argv[0];
195
196	file_setprogname(progname);
197
198
199#ifdef S_IFLNK
200	posixly = getenv("POSIXLY_CORRECT") != NULL;
201	flags |=  posixly ? MAGIC_SYMLINK : 0;
202#endif
203	while ((c = getopt_long(argc, argv, OPTSTRING, long_options,
204	    &longindex)) != -1)
205		switch (c) {
206		case OPT_HELP:
207			help();
208			break;
209		case OPT_APPLE:
210			flags |= MAGIC_APPLE;
211			break;
212		case OPT_EXTENSIONS:
213			flags |= MAGIC_EXTENSION;
214			break;
215		case OPT_MIME_TYPE:
216			flags |= MAGIC_MIME_TYPE;
217			break;
218		case OPT_MIME_ENCODING:
219			flags |= MAGIC_MIME_ENCODING;
220			break;
221		case '0':
222			nulsep++;
223			break;
224		case 'b':
225			bflag++;
226			break;
227		case 'c':
228			action = FILE_CHECK;
229			break;
230		case 'C':
231			action = FILE_COMPILE;
232			break;
233		case 'd':
234			flags |= MAGIC_DEBUG|MAGIC_CHECK;
235			break;
236		case 'E':
237			flags |= MAGIC_ERROR;
238			break;
239		case 'e':
240			for (i = 0; i < __arraycount(nv); i++)
241				if (strcmp(nv[i].name, optarg) == 0)
242					break;
243
244			if (i == __arraycount(nv))
245				errflg++;
246			else
247				flags |= nv[i].value;
248			break;
249
250		case 'f':
251			if(action)
252				usage();
253			if (magic == NULL)
254				if ((magic = load(magicfile, flags)) == NULL)
255					return 1;
256			applyparam(magic);
257			e |= unwrap(magic, optarg);
258			++didsomefiles;
259			break;
260		case 'F':
261			separator = optarg;
262			break;
263		case 'i':
264			flags |= MAGIC_MIME;
265			break;
266		case 'k':
267			flags |= MAGIC_CONTINUE;
268			break;
269		case 'l':
270			action = FILE_LIST;
271			break;
272		case 'm':
273			magicfile = optarg;
274			break;
275		case 'n':
276			++nobuffer;
277			break;
278		case 'N':
279			++nopad;
280			break;
281#if defined(HAVE_UTIME) || defined(HAVE_UTIMES)
282		case 'p':
283			flags |= MAGIC_PRESERVE_ATIME;
284			break;
285#endif
286		case 'P':
287			setparam(optarg);
288			break;
289		case 'r':
290			flags |= MAGIC_RAW;
291			break;
292		case 's':
293			flags |= MAGIC_DEVICES;
294			break;
295		case 'S':
296#ifdef HAVE_LIBSECCOMP
297			sandbox = 0;
298#endif
299			break;
300		case 'v':
301			if (magicfile == NULL)
302				magicfile = magic_getpath(magicfile, action);
303			(void)fprintf(stdout, "%s-%s\n", file_getprogname(),
304			    VERSION);
305			(void)fprintf(stdout, "magic file from %s\n",
306			    magicfile);
307#ifdef HAVE_LIBSECCOMP
308			(void)fprintf(stdout, "seccomp support included\n");
309#endif
310			return 0;
311		case 'z':
312			flags |= MAGIC_COMPRESS;
313			break;
314
315		case 'Z':
316			flags |= MAGIC_COMPRESS|MAGIC_COMPRESS_TRANSP;
317			break;
318#ifdef S_IFLNK
319		case 'L':
320			flags |= MAGIC_SYMLINK;
321			break;
322		case 'h':
323			flags &= ~MAGIC_SYMLINK;
324			break;
325#endif
326		case '?':
327		default:
328			errflg++;
329			break;
330		}
331
332	if (errflg) {
333		usage();
334	}
335	if (e)
336		return e;
337
338#ifdef HAVE_LIBSECCOMP
339#if 0
340	if (sandbox && enable_sandbox_basic() == -1)
341#else
342	if (sandbox && enable_sandbox_full() == -1)
343#endif
344		file_err(EXIT_FAILURE, "SECCOMP initialisation failed");
345#endif /* HAVE_LIBSECCOMP */
346
347	if (MAGIC_VERSION != magic_version())
348		file_warnx("Compiled magic version [%d] "
349		    "does not match with shared library magic version [%d]\n",
350		    MAGIC_VERSION, magic_version());
351
352	switch(action) {
353	case FILE_CHECK:
354	case FILE_COMPILE:
355	case FILE_LIST:
356		/*
357		 * Don't try to check/compile ~/.magic unless we explicitly
358		 * ask for it.
359		 */
360		magic = magic_open(flags|MAGIC_CHECK);
361		if (magic == NULL) {
362			file_warn("Can't create magic");
363			return 1;
364		}
365
366
367		switch(action) {
368		case FILE_CHECK:
369			c = magic_check(magic, magicfile);
370			break;
371		case FILE_COMPILE:
372			c = magic_compile(magic, magicfile);
373			break;
374		case FILE_LIST:
375			c = magic_list(magic, magicfile);
376			break;
377		default:
378			abort();
379		}
380		if (c == -1) {
381			file_warnx("%s", magic_error(magic));
382			e = 1;
383			goto out;
384		}
385		goto out;
386	default:
387		if (magic == NULL)
388			if ((magic = load(magicfile, flags)) == NULL)
389				return 1;
390		applyparam(magic);
391	}
392
393	if (optind == argc) {
394		if (!didsomefiles)
395			usage();
396	}
397	else {
398		size_t j, wid, nw;
399		for (wid = 0, j = CAST(size_t, optind); j < CAST(size_t, argc);
400		    j++) {
401			nw = file_mbswidth(argv[j]);
402			if (nw > wid)
403				wid = nw;
404		}
405		/*
406		 * If bflag is only set twice, set it depending on
407		 * number of files [this is undocumented, and subject to change]
408		 */
409		if (bflag == 2) {
410			bflag = optind >= argc - 1;
411		}
412		for (; optind < argc; optind++)
413			e |= process(magic, argv[optind], wid);
414	}
415
416out:
417	if (magic)
418		magic_close(magic);
419	return e;
420}
421
422private void
423applyparam(magic_t magic)
424{
425	size_t i;
426
427	for (i = 0; i < __arraycount(pm); i++) {
428		if (!pm[i].set)
429			continue;
430		if (magic_setparam(magic, pm[i].tag, &pm[i].value) == -1)
431			file_err(EXIT_FAILURE, "Can't set %s", pm[i].name);
432	}
433}
434
435private void
436setparam(const char *p)
437{
438	size_t i;
439	char *s;
440
441	if ((s = strchr(p, '=')) == NULL)
442		goto badparm;
443
444	for (i = 0; i < __arraycount(pm); i++) {
445		if (strncmp(p, pm[i].name, s - p) != 0)
446			continue;
447		pm[i].value = atoi(s + 1);
448		pm[i].set = 1;
449		return;
450	}
451badparm:
452	file_errx(EXIT_FAILURE, "Unknown param %s", p);
453}
454
455private struct magic_set *
456/*ARGSUSED*/
457load(const char *magicfile, int flags)
458{
459	struct magic_set *magic = magic_open(flags);
460	const char *e;
461
462	if (magic == NULL) {
463		file_warn("Can't create magic");
464		return NULL;
465	}
466	if (magic_load(magic, magicfile) == -1) {
467		file_warn("%s", magic_error(magic));
468		magic_close(magic);
469		return NULL;
470	}
471	if ((e = magic_error(magic)) != NULL)
472		file_warn("%s", e);
473	return magic;
474}
475
476/*
477 * unwrap -- read a file of filenames, do each one.
478 */
479private int
480unwrap(struct magic_set *ms, const char *fn)
481{
482	FILE *f;
483	ssize_t len;
484	char *line = NULL;
485	size_t llen = 0;
486	int wid = 0, cwid;
487	int e = 0;
488
489	if (strcmp("-", fn) == 0) {
490		f = stdin;
491		wid = 1;
492	} else {
493		if ((f = fopen(fn, "r")) == NULL) {
494			file_warn("Cannot open `%s'", fn);
495			return 1;
496		}
497
498		while ((len = getline(&line, &llen, f)) > 0) {
499			if (line[len - 1] == '\n')
500				line[len - 1] = '\0';
501			cwid = file_mbswidth(line);
502			if (cwid > wid)
503				wid = cwid;
504		}
505
506		rewind(f);
507	}
508
509	while ((len = getline(&line, &llen, f)) > 0) {
510		if (line[len - 1] == '\n')
511			line[len - 1] = '\0';
512		e |= process(ms, line, wid);
513		if(nobuffer)
514			(void)fflush(stdout);
515	}
516
517	free(line);
518	(void)fclose(f);
519	return e;
520}
521
522/*
523 * Called for each input file on the command line (or in a list of files)
524 */
525private int
526process(struct magic_set *ms, const char *inname, int wid)
527{
528	const char *type, c = nulsep > 1 ? '\0' : '\n';
529	int std_in = strcmp(inname, "-") == 0;
530
531	if (wid > 0 && !bflag) {
532		(void)printf("%s", std_in ? "/dev/stdin" : inname);
533		if (nulsep)
534			(void)putc('\0', stdout);
535		if (nulsep < 2) {
536			(void)printf("%s", separator);
537			(void)printf("%*s ", CAST(int, nopad ? 0
538			    : (wid - file_mbswidth(inname))), "");
539		}
540	}
541
542	type = magic_file(ms, std_in ? NULL : inname);
543
544	if (type == NULL) {
545		(void)printf("ERROR: %s%c", magic_error(ms), c);
546		return 1;
547	} else {
548		(void)printf("%s%c", type, c);
549		return 0;
550	}
551}
552
553protected size_t
554file_mbswidth(const char *s)
555{
556#if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
557	size_t bytesconsumed, old_n, n, width = 0;
558	mbstate_t state;
559	wchar_t nextchar;
560	(void)memset(&state, 0, sizeof(mbstate_t));
561	old_n = n = strlen(s);
562
563	while (n > 0) {
564		bytesconsumed = mbrtowc(&nextchar, s, n, &state);
565		if (bytesconsumed == CAST(size_t, -1) ||
566		    bytesconsumed == CAST(size_t, -2)) {
567			/* Something went wrong, return something reasonable */
568			return old_n;
569		}
570		if (s[0] == '\n') {
571			/*
572			 * do what strlen() would do, so that caller
573			 * is always right
574			 */
575			width++;
576		} else {
577			int w = wcwidth(nextchar);
578			if (w > 0)
579				width += w;
580		}
581
582		s += bytesconsumed, n -= bytesconsumed;
583	}
584	return width;
585#else
586	return strlen(s);
587#endif
588}
589
590private void
591usage(void)
592{
593	const char *pn = file_getprogname();
594	(void)fprintf(stderr, USAGE, pn, pn, pn);
595	exit(EXIT_FAILURE);
596}
597
598private void
599defprint(int def)
600{
601	if (!def)
602		return;
603	if (((def & 1) && posixly) || ((def & 2) && !posixly))
604		fprintf(stdout, " (default)");
605	fputc('\n', stdout);
606}
607
608private void
609docprint(const char *opts, int def)
610{
611	size_t i;
612	int comma;
613	char *sp, *p;
614
615	p = strstr(opts, "%o");
616	if (p == NULL) {
617		fprintf(stdout, "%s", opts);
618		defprint(def);
619		return;
620	}
621
622	for (sp = p - 1; sp > opts && *sp == ' '; sp--)
623		continue;
624
625	fprintf(stdout, "%.*s", CAST(int, p - opts), opts);
626
627	comma = 0;
628	for (i = 0; i < __arraycount(nv); i++) {
629		fprintf(stdout, "%s%s", comma++ ? ", " : "", nv[i].name);
630		if (i && i % 5 == 0 && i != __arraycount(nv) - 1) {
631			fprintf(stdout, ",\n%*s", CAST(int, p - sp - 1), "");
632			comma = 0;
633		}
634	}
635
636	fprintf(stdout, "%s", opts + (p - opts) + 2);
637}
638
639private void
640help(void)
641{
642	(void)fputs(
643"Usage: file [OPTION...] [FILE...]\n"
644"Determine type of FILEs.\n"
645"\n", stdout);
646#define OPT(shortname, longname, opt, def, doc)      \
647	fprintf(stdout, "  -%c, --" longname, shortname), \
648	docprint(doc, def);
649#define OPT_LONGONLY(longname, opt, def, doc, id)        \
650	fprintf(stdout, "      --" longname),	\
651	docprint(doc, def);
652#include "file_opts.h"
653#undef OPT
654#undef OPT_LONGONLY
655	fprintf(stdout, "\nReport bugs to https://bugs.astron.com/\n");
656	exit(EXIT_SUCCESS);
657}
658
659private const char *file_progname;
660
661protected void
662file_setprogname(const char *progname)
663{
664	file_progname = progname;
665}
666
667protected const char *
668file_getprogname(void)
669{
670	return file_progname;
671}
672
673protected void
674file_err(int e, const char *fmt, ...)
675{
676	va_list ap;
677	int se = errno;
678
679	va_start(ap, fmt);
680	fprintf(stderr, "%s: ", file_progname);
681	vfprintf(stderr, fmt, ap);
682	va_end(ap);
683	fprintf(stderr, " (%s)\n", strerror(se));
684	exit(e);
685}
686
687protected void
688file_errx(int e, const char *fmt, ...)
689{
690	va_list ap;
691
692	va_start(ap, fmt);
693	fprintf(stderr, "%s: ", file_progname);
694	vfprintf(stderr, fmt, ap);
695	va_end(ap);
696	fprintf(stderr, "\n");
697	exit(e);
698}
699
700protected void
701file_warn(const char *fmt, ...)
702{
703	va_list ap;
704	int se = errno;
705
706	va_start(ap, fmt);
707	fprintf(stderr, "%s: ", file_progname);
708	vfprintf(stderr, fmt, ap);
709	va_end(ap);
710	fprintf(stderr, " (%s)\n", strerror(se));
711	errno = se;
712}
713
714protected void
715file_warnx(const char *fmt, ...)
716{
717	va_list ap;
718	int se = errno;
719
720	va_start(ap, fmt);
721	fprintf(stderr, "%s: ", file_progname);
722	vfprintf(stderr, fmt, ap);
723	va_end(ap);
724	fprintf(stderr, "\n");
725	errno = se;
726}
727