1 .\" $FreeBSD: src/usr.bin/file/file.1,v 1.16.2.10 2003/03/16 04:47:04 obrien Exp $
2 .\" $DragonFly: src/usr.bin/file/Attic/file.1,v 1.3 2006/02/17 19:39:04 swildner Exp $
3 .\" $Id: file.man,v 1.44 2003/02/27 20:47:46 christos Exp $
5 .Dt FILE 1 "Copyright but distributable"
9 .Nd determine file type
15 .Op Fl m Ar magicfiles
21 This manual page documents version 3.41 of the
23 utility which tests each argument in an attempt to classify it.
24 There are three sets of tests, performed in this order:
25 file system tests, magic number tests, and language tests.
28 test that succeeds causes the file type to be printed.
30 The type printed will usually contain one of the words
32 (the file contains only
33 printing characters and a few common control
34 characters and is probably safe to read on an
38 (the file contains the result of compiling a program
39 in a form understandable to some
44 meaning anything else (data is usually
47 Exceptions are well-known file formats (core files, tar archives)
48 that are known to contain binary data.
49 When modifying the file
50 .Pa /usr/share/misc/magic
51 or the program itself,
52 .Em "preserve these keywords" .
53 People depend on knowing that all the readable files in a directory
57 Don't do as Berkeley did and change
58 .Dq Li "shell commands text"
60 .Dq Li "shell script" .
62 .Pa /usr/share/misc/magic
63 is built mechanically from a large number of small files in
66 in the source distribution of this program.
68 The file system tests are based on examining the return from a
71 The program checks to see if the file is empty,
72 or if it's some sort of special file.
73 Any known file types appropriate to the system you are running on
74 (sockets, symbolic links, or named pipes (FIFOs) on those systems that
76 are intuited if they are defined in
77 the system header file
80 The magic number tests are used to check for files with data in
81 particular fixed formats.
82 The canonical example of this is a binary executable (compiled program)
84 file, whose format is defined in
88 in the standard include directory.
91 stored in a particular place
92 near the beginning of the file that tells the
95 that the file is a binary executable, and which of several types thereof.
98 has been applied by extension to data files.
99 Any file with some invariant identifier at a small fixed
100 offset into the file can usually be described in this way.
101 The information identifying these files is read from the compiled
103 .Pa /usr/share/misc/magic.mgc ,
105 .Pa /usr/share/misc/magic
106 if the compile file does not exist.
108 If a file does not match any of the entries in the magic file,
109 it is examined to see if it seems to be a text file.
110 ASCII, ISO-8859-x, non-ISO 8-bit extended-ASCII character sets
111 (such as those used on Macintosh and IBM PC systems),
112 UTF-8-encoded Unicode, UTF-16-encoded Unicode, and EBCDIC
113 character sets can be distinguished by the different
114 ranges and sequences of bytes that constitute printable text
116 If a file passes any of these tests, its character set is reported.
117 ASCII, ISO-8859-x, UTF-8, and extended-ASCII files are identified
120 because they will be mostly readable on nearly any terminal;
121 UTF-16 and EBCDIC are only
122 .Dq Li "character data"
124 they contain text, it is text that will require translation
125 before it can be read.
128 will attempt to determine other characteristics of text-type files.
129 If the lines of a file are terminated by CR, CRLF, or NEL, instead
132 LF, this will be reported.
133 Files that contain embedded escape sequences or overstriking
134 will also be identified.
138 has determined the character set used in a text-type file,
140 attempt to determine in what language the file is written.
141 The language tests look for particular strings (cf
143 that can appear anywhere in the first few blocks of a file.
144 For example, the keyword
146 indicates that the file is most likely a
148 input file, just as the keyword
150 indicates a C program.
151 These tests are less reliable than the previous
152 two groups, so they are performed last.
153 The language test routines also test for some miscellany
158 Any file that cannot be identified as having been written
159 in any of the character sets listed above is simply said to be
162 .Bl -tag -width indent
164 Do not prepend filenames to output lines (brief mode).
166 Cause a checking printout of the parsed form of the magic file.
167 This is usually used in conjunction with
169 to debug a new magic file before installing it.
173 output file that contains a pre-parsed version of
176 Read the names of the files to be examined from
179 before the argument list.
182 or at least one filename argument must be present;
183 to test the standard input, use
185 as a filename argument.
186 .It Fl F Ar separator
187 Use the specified separator character instead of
190 Causes the file command to output mime type strings rather than the more
191 traditional human readable ones.
193 .Dq Li "text/plain; charset=us-ascii"
195 .Dq Li "ASCII text" .
196 In order for this option to work, file changes the way
197 it handles files recognised by the command itself (such as many of the
198 text file types, directories etc), and makes use of an alternative
205 Don't stop at the first match, keep going.
207 Specify an alternate list of files containing magic numbers.
208 This can be a single file, or a colon-separated list of files.
210 Force stdout to be flushed after checking each file.
211 This is only useful if checking a list of files.
212 It is intended to be used by programs that want
213 filetype output from a pipe.
215 Don't pad output to align filenames nicely.
217 Print the version of the program and exit.
219 Try to look inside compressed files.
221 option causes symlinks to be followed, as the like-named option in
223 (on systems that support symbolic links).
227 only attempts to read and determine the type of argument files which
229 reports are ordinary files.
230 This prevents problems, because reading special files may have peculiar
236 to also read argument files which are block or character special files.
237 This is useful for determining the file system types of the data in raw
238 disk partitions, which are block special files.
239 This option also causes
241 to disregard the file size as reported by
243 since on some systems it reports a zero size for raw disk partitions.
246 The environment variable
248 can be used to set the default magic number files.
250 .Bl -tag -width ".Pa /usr/share/misc/magic.mime" -compact
251 .It Pa /usr/share/misc/magic.mgc
252 default compiled list of magic numbers
253 .It Pa /usr/share/misc/magic
254 default list of magic numbers
255 .It Pa /usr/share/misc/magic.mime
256 default list of magic numbers, used to output mime types when the
262 $ file file.c file /dev/{wd0a,hda}
263 file.c: C program text
264 file: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV),
265 dynamically linked (uses shared libs), stripped
266 /dev/wd0a: block special (0/0)
267 /dev/hda: block special (3/0)
268 $ file -s /dev/wd0{b,d}
270 /dev/wd0d: x86 boot sector
271 $ file -s /dev/hda{,1,2,3,4,5,6,7,8,9,10} # Linux
272 /dev/hda: x86 boot sector
273 /dev/hda1: Linux/i386 ext2 filesystem
274 /dev/hda2: x86 boot sector
275 /dev/hda3: x86 boot sector, extended partition table
276 /dev/hda4: Linux/i386 ext2 filesystem
277 /dev/hda5: Linux/i386 swap file
278 /dev/hda6: Linux/i386 swap file
279 /dev/hda7: Linux/i386 swap file
280 /dev/hda8: Linux/i386 swap file
284 $ file -s /dev/rwd0e # BSD
286 Unix Fast File system (little-endian),
287 last mounted on /usr,
288 last written at Mon Feb 10 13:22:40 2003,
290 number of blocks 28754208,
291 number of data blocks 27812712,
292 number of cylinder groups 3566,
295 minimum percentage of free blocks 5,
296 rotational delay 0ms,
297 disk rotational speed 60rps,
300 $ file -i file.c file /dev/{wd0a,hda}
302 file: application/x-executable, dynamically linked (uses shared libs),
304 /dev/hda: application/x-not-regular-file
305 /dev/wd0a: application/x-not-regular-file
312 .Sh STANDARDS CONFORMANCE
313 This program is believed to exceed the
315 of FILE(CMD), as near as one can determine from the vague language
317 Its behaviour is mostly compatible with the System V program of the same name.
318 This version knows more magic, however, so it will produce
319 different (albeit more accurate) output in many cases.
321 The one significant difference
322 between this version and System V
323 is that this version treats any white space
324 as a delimiter, so that spaces in pattern strings must be escaped.
327 .Dl ">10 string language impress\ (imPRESS data)"
329 in an existing magic file would have to be changed to
331 .Dl ">10 string language\e impress (imPRESS data)"
333 In addition, in this version, if a pattern string contains a backslash,
337 .Dl "0 string \ebegindata Andrew Toolkit document"
339 in an existing magic file would have to be changed to
341 .Dl "0 string \e\ebegindata Andrew Toolkit document"
343 SunOS releases 3.2 and later from Sun Microsystems include a
345 command derived from the System V one, but with some extensions.
346 My version differs from Sun's only in minor ways.
347 It includes the extension of the
352 .Dl ">16 long&0x7fffffff >0 not stripped"
354 The magic file entries have been collected from various sources,
355 mainly USENET, and contributed by various authors.
357 (address below) will collect additional
358 or corrected magic file entries.
359 A consolidation of magic file entries
360 will be distributed periodically.
362 The order of entries in the magic file is significant.
363 Depending on what system you are using, the order that
364 they are put together may be incorrect.
367 command uses a magic file,
368 keep the old magic file around for comparison purposes
370 .Pa /usr/share/misc/magic.orig ) .
376 since at least Research Version 4
377 (man page dated November, 1973).
378 The System V version introduced one significant major change:
379 the external list of magic number types.
380 This slowed the program down slightly but made it a lot more flexible.
382 This program, based on the System V version,
384 .An Ian Darwin Aq ian@darwinsys.com
385 without looking at anybody else's source code.
388 revised the code extensively, making it better than
391 found several inadequacies
392 and provided some magic file entries.
396 .An Rob McMahon Aq cudcv@warwick.ac.uk ,
399 .An Guy Harris Aq guy@netapp.com ,
400 made many changes from 1993 to the present.
402 Primary development and maintenance from 1990 to the present by
403 .An Christos Zoulas Aq christos@astron.com .
406 .An Chris Lowth Aq chris@lowth.com ,
410 option to output mime type strings and using an alternative
411 magic file and internal logic.
414 .An Eric Fischer Aq enf@pobox.com ,
416 to identify character codes and attempt to identify the languages
421 The list of contributors to the
423 directory (source for the
424 .Pa /usr/share/misc/magic
425 file) is too long to include here.
426 You know who you are; thank you.
428 Copyright (c) Ian F. Darwin, Toronto, Canada, 1986-1999.
429 Covered by the standard Berkeley Software Distribution copyright; see the file
431 in the source distribution.
439 from his public-domain
441 program, and are not covered by the above license.
443 There must be a better way to automate the construction of the
445 file from all the glop in
448 Better yet, the magic file should be compiled into binary (say,
450 or, better yet, fixed-length
452 strings for use in heterogenous network environments) for faster startup.
453 Then the program would run as fast as the Version 7 program of the same name,
454 with the flexibility of the System V version.
458 utility uses several algorithms that favor speed over accuracy,
459 thus it can be misled about the contents of
465 files (primarily for programming languages)
466 is simplistic, inefficient and requires recompilation to update.
470 clause to follow a series of continuation lines.
472 The magic file and keywords should have regular expression support.
475 as a field delimiter is ugly and makes
476 it hard to edit the files, but is entrenched.
478 It might be advisable to allow upper-case letters in keywords
481 commands vs man page macros.
482 Regular expression support would make this easy.
484 The program doesn't grok
486 It should be able to figure
488 by seeing some keywords which
489 appear indented at the start of line.
490 Regular expression support would make this easy.
492 The list of keywords in
494 probably belongs in the
497 This could be done by using some keyword like
499 for the offset value.
501 Another optimisation would be to sort
502 the magic file so that we can just run down all the
503 tests for the first byte, first word, first long, etc, once we
505 Complain about conflicts in the magic file entries.
506 Make a rule that the magic entries sort based on file offset rather
507 than position within the magic file?
509 The program should provide a way to give an estimate
513 We end up removing guesses (e.g.\&
515 as first 5 chars of file) because
516 they are not as good as other guesses (e.g.\&
519 .Dq Li "Return-Path:" ) .
520 Still, if the others don't pan out, it should be
521 possible to use the first guess.
523 This program is slower than some vendors' file commands.
524 The new support for multiple character codes makes it even slower.
526 This manual page, and particularly this section, is too long.
528 You can obtain the original author's latest version by anonymous FTP
532 .Pa /pub/file/file-X.YZ.tar.gz