file.c revision 103373 
1/*
2 * file - find type of a file or files - main program.
3 *
4 * Copyright (c) Ian F. Darwin, 1987.
5 * Written by Ian F. Darwin.
6 *
7 * This software is not subject to any license of the American Telephone
8 * and Telegraph Company or of the Regents of the University of California.
9 *
10 * Permission is granted to anyone to use this software for any purpose on
11 * any computer system, and to alter it and redistribute it freely, subject
12 * to the following restrictions:
13 *
14 * 1. The author is not responsible for the consequences of use of this
15 *    software, no matter how awful, even if they arise from flaws in it.
16 *
17 * 2. The origin of this software must not be misrepresented, either by
18 *    explicit claim or by omission.  Since few users ever read sources,
19 *    credits must appear in the documentation.
20 *
21 * 3. Altered versions must be plainly marked as such, and must not be
22 *    misrepresented as being the original software.  Since few users
23 *    ever read sources, credits must appear in the documentation.
24 *
25 * 4. This notice may not be removed or altered.
26 */
27
28#include "file.h"
29#include <stdlib.h>
30#include <unistd.h>
31#include <string.h>
32#include <sys/param.h>	/* for MAXPATHLEN */
33#include <fcntl.h>	/* for open() */
34#ifdef RESTORE_TIME
35# if (__COHERENT__ >= 0x420)
36#  include <sys/utime.h>
37# else
38#  ifdef USE_UTIMES
39#   include <sys/time.h>
40#  else
41#   include <utime.h>
42#  endif
43# endif
44#endif
45#ifdef HAVE_UNISTD_H
46#include <unistd.h>	/* for read() */
47#endif
48#ifdef HAVE_LOCALE_H
49#include <locale.h>
50#endif
51
52#ifdef HAVE_GETOPT_H
53#include <getopt.h>	/* for long options (is this portable?)*/
54#endif
55
56#include <netinet/in.h>		/* for byte swapping */
57
58#include "patchlevel.h"
59
60#ifndef	lint
61FILE_RCSID("@(#)$Id: file.c,v 1.66 2002/07/03 19:00:41 christos Exp $")
62#endif	/* lint */
63
64
65#ifdef S_IFLNK
66# define USAGE  "Usage: %s [-bciknsvzL] [-f namefile] [-m magicfiles] file...\n"
67#else
68# define USAGE  "Usage: %s [-bciknsvz] [-f namefile] [-m magicfiles] file...\n"
69#endif
70
71#ifdef __EMX__
72static char *apptypeName = NULL;
73int os2_apptype (const char *fn, char *buf, int nb);
74#endif /* __EMX__ */
75
76#ifndef MAGIC
77# define MAGIC "/etc/magic"
78#endif
79
80#ifndef MAXPATHLEN
81#define	MAXPATHLEN	512
82#endif
83
84int 			/* Global command-line options 		*/
85	debug = 0, 	/* debugging 				*/
86	lflag = 0,	/* follow Symlinks (BSD only) 		*/
87	bflag = 0,	/* brief output format	 		*/
88	zflag = 0,	/* follow (uncompress) compressed files */
89	sflag = 0,	/* read block special files		*/
90	iflag = 0,
91	nobuffer = 0,   /* Do not buffer stdout */
92	kflag = 0;	/* Keep going after the first match	*/
93
94int			/* Misc globals				*/
95	nmagic = 0;	/* number of valid magic[]s 		*/
96
97struct  magic *magic;	/* array of magic entries		*/
98
99const char *magicfile = 0;	/* where the magic is		*/
100const char *default_magicfile = MAGIC;
101
102char *progname;		/* used throughout 			*/
103int lineno;		/* line number in the magic file	*/
104
105
106static void	unwrap(char *fn);
107static void	usage(void);
108#ifdef HAVE_GETOPT_H
109static void	help(void);
110#endif
111#if 0
112static int	byteconv4(int, int, int);
113static short	byteconv2(int, int, int);
114#endif
115
116int main(int, char *[]);
117
118/*
119 * main - parse arguments and handle options
120 */
121int
122main(int argc, char **argv)
123{
124	int c;
125	int action = 0, didsomefiles = 0, errflg = 0, ret = 0, app = 0;
126	char *mime, *home, *usermagic;
127	struct stat sb;
128#define OPTSTRING	"bcdf:ikm:nsvzCL"
129#ifdef HAVE_GETOPT_H
130	int longindex;
131	static struct option long_options[] =
132	{
133		{"version", 0, 0, 'v'},
134		{"help", 0, 0, 0},
135		{"brief", 0, 0, 'b'},
136		{"checking-printout", 0, 0, 'c'},
137		{"debug", 0, 0, 'd'},
138		{"files-from", 1, 0, 'f'},
139		{"mime", 0, 0, 'i'},
140		{"keep-going", 0, 0, 'k'},
141#ifdef S_IFLNK
142		{"dereference", 0, 0, 'L'},
143#endif
144		{"magic-file", 1, 0, 'm'},
145		{"uncompress", 0, 0, 'z'},
146		{"no-buffer", 0, 0, 'n'},
147		{"special-files", 0, 0, 's'},
148		{"compile", 0, 0, 'C'},
149		{0, 0, 0, 0},
150	};
151#endif
152
153#ifdef LC_CTYPE
154	setlocale(LC_CTYPE, ""); /* makes islower etc work for other langs */
155#endif
156
157#ifdef __EMX__
158	/* sh-like wildcard expansion! Shouldn't hurt at least ... */
159	_wildcard(&argc, &argv);
160#endif
161
162	if ((progname = strrchr(argv[0], '/')) != NULL)
163		progname++;
164	else
165		progname = argv[0];
166
167	magicfile = default_magicfile;
168	if ((usermagic = getenv("MAGIC")) != NULL)
169		magicfile = usermagic;
170	else
171		if ((home = getenv("HOME")) != NULL) {
172			if ((usermagic = malloc(strlen(home) + 8)) != NULL) {
173				(void)strcpy(usermagic, home);
174				(void)strcat(usermagic, "/.magic");
175				if (stat(usermagic, &sb)<0)
176					free(usermagic);
177				else
178					magicfile = usermagic;
179			}
180		}
181
182#ifndef HAVE_GETOPT_H
183	while ((c = getopt(argc, argv, OPTSTRING)) != -1)
184#else
185	while ((c = getopt_long(argc, argv, OPTSTRING, long_options,
186	    &longindex)) != -1)
187#endif
188		switch (c) {
189#ifdef HAVE_GETOPT_H
190		case 0 :
191			if (longindex == 1)
192				help();
193			break;
194#endif
195		case 'b':
196			++bflag;
197			break;
198		case 'c':
199			action = CHECK;
200			break;
201		case 'C':
202			action = COMPILE;
203			break;
204		case 'd':
205			++debug;
206			break;
207		case 'f':
208			if (!app) {
209				ret = apprentice(magicfile, action);
210				if (action)
211					exit(ret);
212				app = 1;
213			}
214			unwrap(optarg);
215			++didsomefiles;
216			break;
217		case 'i':
218			iflag++;
219			if ((mime = malloc(strlen(magicfile) + 6)) != NULL) {
220				(void)strcpy(mime, magicfile);
221				(void)strcat(mime, ".mime");
222				magicfile = mime;
223			}
224			break;
225		case 'k':
226			kflag = 1;
227			break;
228		case 'm':
229			magicfile = optarg;
230			break;
231		case 'n':
232			++nobuffer;
233			break;
234		case 's':
235			sflag++;
236			break;
237		case 'v':
238			(void) fprintf(stdout, "%s-%d.%d\n", progname,
239				       FILE_VERSION_MAJOR, patchlevel);
240			(void) fprintf(stdout, "magic file from %s\n",
241				       magicfile);
242			return 1;
243		case 'z':
244			zflag++;
245			break;
246#ifdef S_IFLNK
247		case 'L':
248			++lflag;
249			break;
250#endif
251		case '?':
252		default:
253			errflg++;
254			break;
255		}
256
257	if (errflg) {
258		usage();
259	}
260
261	if (!app) {
262		ret = apprentice(magicfile, action);
263		if (action)
264			exit(ret);
265		app = 1;
266	}
267
268	if (optind == argc) {
269		if (!didsomefiles) {
270			usage();
271		}
272	}
273	else {
274		int i, wid, nw;
275		for (wid = 0, i = optind; i < argc; i++) {
276			nw = strlen(argv[i]);
277			if (nw > wid)
278				wid = nw;
279		}
280		for (; optind < argc; optind++)
281			process(argv[optind], wid);
282	}
283
284	return 0;
285}
286
287
288/*
289 * unwrap -- read a file of filenames, do each one.
290 */
291static void
292unwrap(char *fn)
293{
294	char buf[MAXPATHLEN];
295	FILE *f;
296	int wid = 0, cwid;
297
298	if (strcmp("-", fn) == 0) {
299		f = stdin;
300		wid = 1;
301	} else {
302		if ((f = fopen(fn, "r")) == NULL) {
303			error("Cannot open `%s' (%s).\n", fn, strerror(errno));
304			/*NOTREACHED*/
305		}
306
307		while (fgets(buf, MAXPATHLEN, f) != NULL) {
308			cwid = strlen(buf) - 1;
309			if (cwid > wid)
310				wid = cwid;
311		}
312
313		rewind(f);
314	}
315
316	while (fgets(buf, MAXPATHLEN, f) != NULL) {
317		buf[strlen(buf)-1] = '\0';
318		process(buf, wid);
319		if(nobuffer)
320			(void) fflush(stdout);
321	}
322
323	(void) fclose(f);
324}
325
326
327#if 0
328/*
329 * byteconv4
330 * Input:
331 *	from		4 byte quantity to convert
332 *	same		whether to perform byte swapping
333 *	big_endian	whether we are a big endian host
334 */
335static int
336byteconv4(int from, int same, int big_endian)
337{
338	if (same)
339		return from;
340	else if (big_endian) {		/* lsb -> msb conversion on msb */
341		union {
342			int i;
343			char c[4];
344		} retval, tmpval;
345
346		tmpval.i = from;
347		retval.c[0] = tmpval.c[3];
348		retval.c[1] = tmpval.c[2];
349		retval.c[2] = tmpval.c[1];
350		retval.c[3] = tmpval.c[0];
351
352		return retval.i;
353	}
354	else
355		return ntohl(from);	/* msb -> lsb conversion on lsb */
356}
357
358/*
359 * byteconv2
360 * Same as byteconv4, but for shorts
361 */
362static short
363byteconv2(int from, int same, int big_endian)
364{
365	if (same)
366		return from;
367	else if (big_endian) {		/* lsb -> msb conversion on msb */
368		union {
369			short s;
370			char c[2];
371		} retval, tmpval;
372
373		tmpval.s = (short) from;
374		retval.c[0] = tmpval.c[1];
375		retval.c[1] = tmpval.c[0];
376
377		return retval.s;
378	}
379	else
380		return ntohs(from);	/* msb -> lsb conversion on lsb */
381}
382#endif
383
384/*
385 * process - process input file
386 */
387void
388process(const char *inname, int wid)
389{
390	int	fd = 0;
391	static  const char stdname[] = "standard input";
392	unsigned char	buf[HOWMANY+1];	/* one extra for terminating '\0' */
393	struct stat	sb;
394	int nbytes = 0;	/* number of bytes read from a datafile */
395	char match = '\0';
396
397	if (strcmp("-", inname) == 0) {
398		if (fstat(0, &sb)<0) {
399			error("cannot fstat `%s' (%s).\n", stdname,
400			      strerror(errno));
401			/*NOTREACHED*/
402		}
403		inname = stdname;
404	}
405
406	if (wid > 0 && !bflag)
407	     (void) printf("%s:%*s ", inname,
408			   (int) (wid - strlen(inname)), "");
409
410	if (inname != stdname) {
411		/*
412		 * first try judging the file based on its filesystem status
413		 */
414		if (fsmagic(inname, &sb) != 0) {
415			putchar('\n');
416			return;
417		}
418
419		if ((fd = open(inname, O_RDONLY)) < 0) {
420			/* We can't open it, but we were able to stat it. */
421			if (sb.st_mode & 0002) ckfputs("writeable, ", stdout);
422			if (sb.st_mode & 0111) ckfputs("executable, ", stdout);
423			ckfprintf(stdout, "can't read `%s' (%s).\n",
424			    inname, strerror(errno));
425			return;
426		}
427	}
428
429
430	/*
431	 * try looking at the first HOWMANY bytes
432	 */
433	if ((nbytes = read(fd, (char *)buf, HOWMANY)) == -1) {
434		error("read failed (%s).\n", strerror(errno));
435		/*NOTREACHED*/
436	}
437
438	if (nbytes == 0)
439		ckfputs(iflag ? "application/x-empty" : "empty", stdout);
440	else {
441		buf[nbytes++] = '\0';	/* null-terminate it */
442		match = tryit(inname, buf, nbytes, zflag);
443	}
444
445#ifdef BUILTIN_ELF
446	if (match == 's' && nbytes > 5) {
447		/*
448		 * We matched something in the file, so this *might*
449		 * be an ELF file, and the file is at least 5 bytes long,
450		 * so if it's an ELF file it has at least one byte
451		 * past the ELF magic number - try extracting information
452		 * from the ELF headers that can't easily be extracted
453		 * with rules in the magic file.
454		 */
455		tryelf(fd, buf, nbytes);
456	}
457#endif
458
459	if (inname != stdname) {
460#ifdef RESTORE_TIME
461		/*
462		 * Try to restore access, modification times if read it.
463		 * This is really *bad* because it will modify the status
464		 * time of the file... And of course this will affect
465		 * backup programs
466		 */
467# ifdef USE_UTIMES
468		struct timeval  utsbuf[2];
469		utsbuf[0].tv_sec = sb.st_atime;
470		utsbuf[1].tv_sec = sb.st_mtime;
471
472		(void) utimes(inname, utsbuf); /* don't care if loses */
473# else
474		struct utimbuf  utbuf;
475
476		utbuf.actime = sb.st_atime;
477		utbuf.modtime = sb.st_mtime;
478		(void) utime(inname, &utbuf); /* don't care if loses */
479# endif
480#endif
481		(void) close(fd);
482	}
483	(void) putchar('\n');
484}
485
486
487int
488tryit(const char *fn, unsigned char *buf, int nb, int zfl)
489{
490
491	/*
492	 * The main work is done here!
493	 * We have the file name and/or the data buffer to be identified.
494	 */
495
496#ifdef __EMX__
497	/*
498	 * Ok, here's the right place to add a call to some os-specific
499	 * routine, e.g.
500	 */
501	if (os2_apptype(fn, buf, nb) == 1)
502	       return 'o';
503#endif
504	/* try compression stuff */
505	if (zfl && zmagic(fn, buf, nb))
506		return 'z';
507
508	/* try tests in /etc/magic (or surrogate magic file) */
509	if (softmagic(buf, nb))
510		return 's';
511
512	/* try known keywords, check whether it is ASCII */
513	if (ascmagic(buf, nb))
514		return 'a';
515
516	/* abandon hope, all ye who remain here */
517	ckfputs(iflag ? "application/octet-stream" : "data", stdout);
518		return '\0';
519}
520
521static void
522usage(void)
523{
524	(void)fprintf(stderr, USAGE, progname);
525	(void)fprintf(stderr, "Usage: %s -C [-m magic]\n", progname);
526#ifdef HAVE_GETOPT_H
527	(void)fputs("Try `file --help' for more information.\n", stderr);
528#endif
529	exit(1);
530}
531
532#ifdef HAVE_GETOPT_H
533static void
534help(void)
535{
536	puts(
537"Usage: file [OPTION]... [FILE]...\n"
538"Determine file type of FILEs.\n"
539"\n"
540"  -m, --magic-file LIST      use LIST as a colon-separated list of magic\n"
541"                               number files\n"
542"  -z, --uncompress           try to look inside compressed files\n"
543"  -b, --brief                do not prepend filenames to output lines\n"
544"  -c, --checking-printout    print the parsed form of the magic file, use in\n"
545"                               conjunction with -m to debug a new magic file\n"
546"                               before installing it\n"
547"  -f, --files-from FILE      read the filenames to be examined from FILE\n"
548"  -i, --mime                 output mime type strings\n"
549"  -k, --keep-going           don't stop at the first match\n"
550"  -L, --dereference          causes symlinks to be followed\n"
551"  -n, --no-buffer            do not buffer output\n"
552"  -s, --special-files        treat special (block/char devices) files as\n"
553"                             ordinary ones\n"
554"      --help                 display this help and exit\n"
555"      --version              output version information and exit\n"
556);
557	exit(0);
558}
559#endif
560