file.c revision 110949 
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 1.68 2003/02/08 18:33:53 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	nopad = 0,	/* Don't pad output			*/
92	nobuffer = 0,   /* Do not buffer stdout 		*/
93	kflag = 0;	/* Keep going after the first match	*/
94
95int			/* Misc globals				*/
96	nmagic = 0;	/* number of valid magic[]s 		*/
97
98struct  magic *magic;	/* array of magic entries		*/
99
100const char *magicfile = 0;	/* where the magic is		*/
101const char *default_magicfile = MAGIC;
102
103char separator = ':';	/* Default field separator		*/
104
105char *progname;		/* used throughout 			*/
106int lineno;		/* line number in the magic file	*/
107
108
109static void	unwrap(char *fn);
110static void	usage(void);
111#ifdef HAVE_GETOPT_LONG
112static void	help(void);
113#endif
114#if 0
115static int	byteconv4(int, int, int);
116static short	byteconv2(int, int, int);
117#endif
118
119int main(int, char *[]);
120
121/*
122 * main - parse arguments and handle options
123 */
124int
125main(int argc, char **argv)
126{
127	int c;
128	int action = 0, didsomefiles = 0, errflg = 0, ret = 0, app = 0;
129	char *mime, *home, *usermagic;
130	struct stat sb;
131#define OPTSTRING	"bcdf:F:ikm:nNsvzCL"
132#ifdef HAVE_GETOPT_LONG
133	int longindex;
134	static struct option long_options[] =
135	{
136		{"version", 0, 0, 'v'},
137		{"help", 0, 0, 0},
138		{"brief", 0, 0, 'b'},
139		{"checking-printout", 0, 0, 'c'},
140		{"debug", 0, 0, 'd'},
141		{"files-from", 1, 0, 'f'},
142		{"separator", 1, 0, 'F'},
143		{"mime", 0, 0, 'i'},
144		{"keep-going", 0, 0, 'k'},
145#ifdef S_IFLNK
146		{"dereference", 0, 0, 'L'},
147#endif
148		{"magic-file", 1, 0, 'm'},
149		{"uncompress", 0, 0, 'z'},
150		{"no-buffer", 0, 0, 'n'},
151		{"no-pad", 0, 0, 'N'},
152		{"special-files", 0, 0, 's'},
153		{"compile", 0, 0, 'C'},
154		{0, 0, 0, 0},
155	};
156#endif
157
158#ifdef LC_CTYPE
159	setlocale(LC_CTYPE, ""); /* makes islower etc work for other langs */
160#endif
161
162#ifdef __EMX__
163	/* sh-like wildcard expansion! Shouldn't hurt at least ... */
164	_wildcard(&argc, &argv);
165#endif
166
167	if ((progname = strrchr(argv[0], '/')) != NULL)
168		progname++;
169	else
170		progname = argv[0];
171
172	magicfile = default_magicfile;
173	if ((usermagic = getenv("MAGIC")) != NULL)
174		magicfile = usermagic;
175	else
176		if ((home = getenv("HOME")) != NULL) {
177			if ((usermagic = malloc(strlen(home) + 8)) != NULL) {
178				(void)strcpy(usermagic, home);
179				(void)strcat(usermagic, "/.magic");
180				if (stat(usermagic, &sb)<0)
181					free(usermagic);
182				else
183					magicfile = usermagic;
184			}
185		}
186
187#ifndef HAVE_GETOPT_LONG
188	while ((c = getopt(argc, argv, OPTSTRING)) != -1)
189#else
190	while ((c = getopt_long(argc, argv, OPTSTRING, long_options,
191	    &longindex)) != -1)
192#endif
193		switch (c) {
194#ifdef HAVE_GETOPT_LONG
195		case 0 :
196			if (longindex == 1)
197				help();
198			break;
199#endif
200		case 'b':
201			++bflag;
202			break;
203		case 'c':
204			action = CHECK;
205			break;
206		case 'C':
207			action = COMPILE;
208			break;
209		case 'd':
210			++debug;
211			break;
212		case 'f':
213			if (!app) {
214				ret = apprentice(magicfile, action);
215				if (action)
216					exit(ret);
217				app = 1;
218			}
219			unwrap(optarg);
220			++didsomefiles;
221			break;
222		case 'F':
223			separator = *optarg;
224			break;
225		case 'i':
226			iflag++;
227			if ((mime = malloc(strlen(magicfile) + 6)) != NULL) {
228				(void)strcpy(mime, magicfile);
229				(void)strcat(mime, ".mime");
230				magicfile = mime;
231			}
232			break;
233		case 'k':
234			kflag = 1;
235			break;
236		case 'm':
237			magicfile = optarg;
238			break;
239		case 'n':
240			++nobuffer;
241			break;
242		case 'N':
243			++nopad;
244			break;
245		case 's':
246			sflag++;
247			break;
248		case 'v':
249			(void) fprintf(stdout, "%s-%d.%d\n", progname,
250				       FILE_VERSION_MAJOR, patchlevel);
251			(void) fprintf(stdout, "magic file from %s\n",
252				       magicfile);
253			return 1;
254		case 'z':
255			zflag++;
256			break;
257#ifdef S_IFLNK
258		case 'L':
259			++lflag;
260			break;
261#endif
262		case '?':
263		default:
264			errflg++;
265			break;
266		}
267
268	if (errflg) {
269		usage();
270	}
271
272	if (!app) {
273		ret = apprentice(magicfile, action);
274		if (action)
275			exit(ret);
276		app = 1;
277	}
278
279	if (optind == argc) {
280		if (!didsomefiles) {
281			usage();
282		}
283	}
284	else {
285		int i, wid, nw;
286		for (wid = 0, i = optind; i < argc; i++) {
287			nw = strlen(argv[i]);
288			if (nw > wid)
289				wid = nw;
290		}
291		for (; optind < argc; optind++)
292			process(argv[optind], wid);
293	}
294
295	return 0;
296}
297
298
299/*
300 * unwrap -- read a file of filenames, do each one.
301 */
302static void
303unwrap(char *fn)
304{
305	char buf[MAXPATHLEN];
306	FILE *f;
307	int wid = 0, cwid;
308
309	if (strcmp("-", fn) == 0) {
310		f = stdin;
311		wid = 1;
312	} else {
313		if ((f = fopen(fn, "r")) == NULL) {
314			error("Cannot open `%s' (%s).\n", fn, strerror(errno));
315			/*NOTREACHED*/
316		}
317
318		while (fgets(buf, MAXPATHLEN, f) != NULL) {
319			cwid = strlen(buf) - 1;
320			if (cwid > wid)
321				wid = cwid;
322		}
323
324		rewind(f);
325	}
326
327	while (fgets(buf, MAXPATHLEN, f) != NULL) {
328		buf[strlen(buf)-1] = '\0';
329		process(buf, wid);
330		if(nobuffer)
331			(void) fflush(stdout);
332	}
333
334	(void) fclose(f);
335}
336
337
338#if 0
339/*
340 * byteconv4
341 * Input:
342 *	from		4 byte quantity to convert
343 *	same		whether to perform byte swapping
344 *	big_endian	whether we are a big endian host
345 */
346static int
347byteconv4(int from, int same, int big_endian)
348{
349	if (same)
350		return from;
351	else if (big_endian) {		/* lsb -> msb conversion on msb */
352		union {
353			int i;
354			char c[4];
355		} retval, tmpval;
356
357		tmpval.i = from;
358		retval.c[0] = tmpval.c[3];
359		retval.c[1] = tmpval.c[2];
360		retval.c[2] = tmpval.c[1];
361		retval.c[3] = tmpval.c[0];
362
363		return retval.i;
364	}
365	else
366		return ntohl(from);	/* msb -> lsb conversion on lsb */
367}
368
369/*
370 * byteconv2
371 * Same as byteconv4, but for shorts
372 */
373static short
374byteconv2(int from, int same, int big_endian)
375{
376	if (same)
377		return from;
378	else if (big_endian) {		/* lsb -> msb conversion on msb */
379		union {
380			short s;
381			char c[2];
382		} retval, tmpval;
383
384		tmpval.s = (short) from;
385		retval.c[0] = tmpval.c[1];
386		retval.c[1] = tmpval.c[0];
387
388		return retval.s;
389	}
390	else
391		return ntohs(from);	/* msb -> lsb conversion on lsb */
392}
393#endif
394
395/*
396 * process - process input file
397 */
398void
399process(const char *inname, int wid)
400{
401	int	fd = 0;
402	static  const char stdname[] = "standard input";
403	unsigned char	buf[HOWMANY+1];	/* one extra for terminating '\0' */
404	struct stat	sb;
405	int nbytes = 0;	/* number of bytes read from a datafile */
406	char match = '\0';
407
408	if (strcmp("-", inname) == 0) {
409		if (fstat(0, &sb)<0) {
410			error("cannot fstat `%s' (%s).\n", stdname,
411			      strerror(errno));
412			/*NOTREACHED*/
413		}
414		inname = stdname;
415	}
416
417	if (wid > 0 && !bflag)
418	     (void) printf("%s%c%*s", inname, separator,
419			   (int) (nopad ? 0 : 1 + (wid - strlen(inname))), "");
420
421	if (inname != stdname) {
422		/*
423		 * first try judging the file based on its filesystem status
424		 */
425		if (fsmagic(inname, &sb) != 0) {
426			putchar('\n');
427			return;
428		}
429
430		if ((fd = open(inname, O_RDONLY)) < 0) {
431			/* We can't open it, but we were able to stat it. */
432			if (sb.st_mode & 0002) ckfputs("writeable, ", stdout);
433			if (sb.st_mode & 0111) ckfputs("executable, ", stdout);
434			ckfprintf(stdout, "can't read `%s' (%s).\n",
435			    inname, strerror(errno));
436			return;
437		}
438	}
439
440
441	/*
442	 * try looking at the first HOWMANY bytes
443	 */
444	if ((nbytes = read(fd, (char *)buf, HOWMANY)) == -1) {
445		error("read failed (%s).\n", strerror(errno));
446		/*NOTREACHED*/
447	}
448
449	if (nbytes == 0)
450		ckfputs(iflag ? "application/x-empty" : "empty", stdout);
451	else {
452		buf[nbytes++] = '\0';	/* null-terminate it */
453		match = tryit(inname, buf, nbytes, zflag);
454	}
455
456#ifdef BUILTIN_ELF
457	if (match == 's' && nbytes > 5) {
458		/*
459		 * We matched something in the file, so this *might*
460		 * be an ELF file, and the file is at least 5 bytes long,
461		 * so if it's an ELF file it has at least one byte
462		 * past the ELF magic number - try extracting information
463		 * from the ELF headers that can't easily be extracted
464		 * with rules in the magic file.
465		 */
466		tryelf(fd, buf, nbytes);
467	}
468#endif
469
470	if (inname != stdname) {
471#ifdef RESTORE_TIME
472		/*
473		 * Try to restore access, modification times if read it.
474		 * This is really *bad* because it will modify the status
475		 * time of the file... And of course this will affect
476		 * backup programs
477		 */
478# ifdef USE_UTIMES
479		struct timeval  utsbuf[2];
480		utsbuf[0].tv_sec = sb.st_atime;
481		utsbuf[1].tv_sec = sb.st_mtime;
482
483		(void) utimes(inname, utsbuf); /* don't care if loses */
484# else
485		struct utimbuf  utbuf;
486
487		utbuf.actime = sb.st_atime;
488		utbuf.modtime = sb.st_mtime;
489		(void) utime(inname, &utbuf); /* don't care if loses */
490# endif
491#endif
492		(void) close(fd);
493	}
494	(void) putchar('\n');
495}
496
497
498int
499tryit(const char *fn, unsigned char *buf, int nb, int zfl)
500{
501
502	/*
503	 * The main work is done here!
504	 * We have the file name and/or the data buffer to be identified.
505	 */
506
507#ifdef __EMX__
508	/*
509	 * Ok, here's the right place to add a call to some os-specific
510	 * routine, e.g.
511	 */
512	if (os2_apptype(fn, buf, nb) == 1)
513	       return 'o';
514#endif
515	/* try compression stuff */
516	if (zfl && zmagic(fn, buf, nb))
517		return 'z';
518
519	/* try tests in /etc/magic (or surrogate magic file) */
520	if (softmagic(buf, nb))
521		return 's';
522
523	/* try known keywords, check whether it is ASCII */
524	if (ascmagic(buf, nb))
525		return 'a';
526
527	/* abandon hope, all ye who remain here */
528	ckfputs(iflag ? "application/octet-stream" : "data", stdout);
529		return '\0';
530}
531
532static void
533usage(void)
534{
535	(void)fprintf(stderr, USAGE, progname);
536	(void)fprintf(stderr, "Usage: %s -C [-m magic]\n", progname);
537#ifdef HAVE_GETOPT_LONG
538	(void)fputs("Try `file --help' for more information.\n", stderr);
539#endif
540	exit(1);
541}
542
543#ifdef HAVE_GETOPT_LONG
544static void
545help(void)
546{
547	puts(
548"Usage: file [OPTION]... [FILE]...\n"
549"Determine file type of FILEs.\n"
550"\n"
551"  -m, --magic-file LIST      use LIST as a colon-separated list of magic\n"
552"                               number files\n"
553"  -z, --uncompress           try to look inside compressed files\n"
554"  -b, --brief                do not prepend filenames to output lines\n"
555"  -c, --checking-printout    print the parsed form of the magic file, use in\n"
556"                               conjunction with -m to debug a new magic file\n"
557"                               before installing it\n"
558"  -f, --files-from FILE      read the filenames to be examined from FILE\n"
559"  -i, --mime                 output mime type strings\n"
560"  -k, --keep-going           don't stop at the first match\n"
561"  -L, --dereference          causes symlinks to be followed\n"
562"  -n, --no-buffer            do not buffer output\n"
563"  -s, --special-files        treat special (block/char devices) files as\n"
564"                             ordinary ones\n"
565"      --help                 display this help and exit\n"
566"      --version              output version information and exit\n"
567);
568	exit(0);
569}
570#endif
571