1 .\" $File: file.man,v 1.144 2021/02/05 22:08:31 christos Exp $
7 .Nd determine file type
11 .Op Fl bcdEhiklLNnprsSvzZ0
13 .Op Fl Fl exclude-quiet
15 .Op Fl Fl mime-encoding
20 .Op Fl m Ar magicfiles
21 .Op Fl P Ar name=value
26 .Op Fl m Ar magicfiles
30 This manual page documents version __VERSION__ of the
35 tests each argument in an attempt to classify it.
36 There are three sets of tests, performed in this order:
37 filesystem tests, magic tests, and language tests.
40 test that succeeds causes the file type to be printed.
42 The type printed will usually contain one of the words
44 (the file contains only
45 printing characters and a few common control
46 characters and is probably safe to read on an
50 (the file contains the result of compiling a program
51 in a form understandable to some
56 meaning anything else (data is usually
59 Exceptions are well-known file formats (core files, tar archives)
60 that are known to contain binary data.
61 When modifying magic files or the program itself, make sure to
62 .Em preserve these keywords .
63 Users depend on knowing that all the readable files in a directory
67 Don't do as Berkeley did and change
68 .Dq shell commands text
72 The filesystem tests are based on examining the return from a
75 The program checks to see if the file is empty,
76 or if it's some sort of special file.
77 Any known file types appropriate to the system you are running on
78 (sockets, symbolic links, or named pipes (FIFOs) on those systems that
80 are intuited if they are defined in the system header file
83 The magic tests are used to check for files with data in
84 particular fixed formats.
85 The canonical example of this is a binary executable (compiled program)
87 file, whose format is defined in
92 in the standard include directory.
95 stored in a particular place
96 near the beginning of the file that tells the
99 that the file is a binary executable, and which of several types thereof.
102 has been applied by extension to data files.
103 Any file with some invariant identifier at a small fixed
104 offset into the file can usually be described in this way.
105 The information identifying these files is read from the compiled
108 or the files in the directory
110 if the compiled file does not exist.
115 exists, it will be used in preference to the system magic files.
117 If a file does not match any of the entries in the magic file,
118 it is examined to see if it seems to be a text file.
119 ASCII, ISO-8859-x, non-ISO 8-bit extended-ASCII character sets
120 (such as those used on Macintosh and IBM PC systems),
121 UTF-8-encoded Unicode, UTF-16-encoded Unicode, and EBCDIC
122 character sets can be distinguished by the different
123 ranges and sequences of bytes that constitute printable text
125 If a file passes any of these tests, its character set is reported.
126 ASCII, ISO-8859-x, UTF-8, and extended-ASCII files are identified
129 because they will be mostly readable on nearly any terminal;
130 UTF-16 and EBCDIC are only
133 they contain text, it is text that will require translation
134 before it can be read.
137 will attempt to determine other characteristics of text-type files.
138 If the lines of a file are terminated by CR, CRLF, or NEL, instead
139 of the Unix-standard LF, this will be reported.
140 Files that contain embedded escape sequences or overstriking
141 will also be identified.
145 has determined the character set used in a text-type file,
147 attempt to determine in what language the file is written.
148 The language tests look for particular strings (cf.
150 that can appear anywhere in the first few blocks of a file.
151 For example, the keyword
153 indicates that the file is most likely a
155 input file, just as the keyword
157 indicates a C program.
158 These tests are less reliable than the previous
159 two groups, so they are performed last.
160 The language test routines also test for some miscellany
163 archives, JSON files).
165 Any file that cannot be identified as having been written
166 in any of the character sets listed above is simply said to be
169 .Bl -tag -width indent
173 command to output the file type and creator code as
174 used by older MacOS versions.
175 The code consists of eight letters,
176 the first describing the file type, the latter the creator.
177 This option works properly only for file formats that have the
178 apple-style output defined.
179 .It Fl b , Fl Fl brief
180 Do not prepend filenames to output lines (brief mode).
181 .It Fl C , Fl Fl compile
184 output file that contains a pre-parsed version of the magic file or directory.
185 .It Fl c , Fl Fl checking-printout
186 Cause a checking printout of the parsed form of the magic file.
187 This is usually used in conjunction with the
189 option to debug a new magic file before installing it.
191 Prints internal debugging information to stderr.
193 On filesystem errors (file not found etc), instead of handling the error
194 as regular output as POSIX mandates and keep going, issue an error message
196 .It Fl e , Fl Fl exclude Ar testname
197 Exclude the test named in
199 from the list of tests made to determine the file type.
200 Valid test names are:
201 .Bl -tag -width compress
204 application type (only on EMX).
206 Various types of text files (this test will try to guess the text
207 encoding, irrespective of the setting of the
211 Different text encodings for soft magic tests.
213 Ignored for backwards compatibility.
215 Prints details of Compound Document Files.
217 Checks for, and looks inside, compressed files.
219 Checks Comma Separated Value files.
221 Prints ELF file details, provided soft magic tests are enabled and the
224 Examines JSON (RFC-7159) files by parsing them for compliance.
226 Consults magic files.
228 Examines tar files by verifying the checksum of the 512 byte tar header.
229 Excluding this test can provide more detailed content description by using
230 the soft magic method.
235 .It Fl Fl exclude-quiet
238 but ignore tests that
241 This is intended for compatibility with older versions of
244 Print a slash-separated list of valid extensions for the file type found.
245 .It Fl F , Fl Fl separator Ar separator
246 Use the specified string as the separator between the filename and the
247 file result returned.
250 .It Fl f , Fl Fl files-from Ar namefile
251 Read the names of the files to be examined from
254 before the argument list.
257 or at least one filename argument must be present;
258 to test the standard input, use
260 as a filename argument.
263 is unwrapped and the enclosed filenames are processed when this option is
264 encountered and before any further options processing is done.
265 This allows one to process multiple lists of files with different command line
266 arguments on the same
269 Thus if you want to set the delimiter, you need to do it before you specify
270 the list of files, like:
271 .Dq Fl F Ar @ Fl f Ar namefile ,
273 .Dq Fl f Ar namefile Fl F Ar @ .
274 .It Fl h , Fl Fl no-dereference
275 This option causes symlinks not to be followed
276 (on systems that support symbolic links).
277 This is the default if the environment variable
280 .It Fl i , Fl Fl mime
283 command to output mime type strings rather than the more
284 traditional human readable ones.
286 .Sq text/plain; charset=us-ascii
289 .It Fl Fl mime-type , Fl Fl mime-encoding
292 but print only the specified element(s).
293 .It Fl k , Fl Fl keep-going
294 Don't stop at the first match, keep going.
295 Subsequent matches will be
299 (If you want a newline, see the
302 The magic pattern with the highest strength (see the
305 .It Fl l , Fl Fl list
306 Shows a list of patterns and their strength sorted descending by
307 .Xr magic __FSECTION__
309 which is used for the matching (see also the
312 .It Fl L , Fl Fl dereference
313 This option causes symlinks to be followed, as the like-named option in
315 (on systems that support symbolic links).
316 This is the default if the environment variable
319 .It Fl m , Fl Fl magic-file Ar magicfiles
320 Specify an alternate list of files and directories containing magic.
321 This can be a single item, or a colon-separated list.
322 If a compiled magic file is found alongside a file or directory,
323 it will be used instead.
324 .It Fl N , Fl Fl no-pad
325 Don't pad filenames so that they align in the output.
326 .It Fl n , Fl Fl no-buffer
327 Force stdout to be flushed after checking each file.
328 This is only useful if checking a list of files.
329 It is intended to be used by programs that want filetype output from a pipe.
330 .It Fl p , Fl Fl preserve-date
331 On systems that support
335 attempt to preserve the access time of files analyzed, to pretend that
338 .It Fl P , Fl Fl parameter Ar name=value
339 Set various parameter limits.
340 .Bl -column "elf_phnum" "Default" "XXXXXXXXXXXXXXXXXXXXXXXXXXX" -offset indent
341 .It Sy "Name" Ta Sy "Default" Ta Sy "Explanation"
342 .It Li bytes Ta 1048576 Ta max number of bytes to read from file
343 .It Li elf_notes Ta 256 Ta max ELF notes processed
344 .It Li elf_phnum Ta 2048 Ta max ELF program sections processed
345 .It Li elf_shnum Ta 32768 Ta max ELF sections processed
346 .It Li encoding Ta 65536 Ta max number of bytes to scan for encoding evaluation
347 .It Li indir Ta 50 Ta recursion limit for indirect magic
348 .It Li name Ta 50 Ta use count limit for name/use magic
349 .It Li regex Ta 8192 Ta length limit for regex searches
352 Don't translate unprintable characters to \eooo.
355 translates unprintable characters to their octal representation.
356 .It Fl s , Fl Fl special-files
359 only attempts to read and determine the type of argument files which
361 reports are ordinary files.
362 This prevents problems, because reading special files may have peculiar
368 to also read argument files which are block or character special files.
369 This is useful for determining the filesystem types of the data in raw
370 disk partitions, which are block special files.
371 This option also causes
373 to disregard the file size as reported by
375 since on some systems it reports a zero size for raw disk partitions.
376 .It Fl S , Fl Fl no-sandbox
377 On systems where libseccomp
378 .Pa ( https://github.com/seccomp/libseccomp )
381 option disables sandboxing which is enabled by default.
382 This option is needed for
384 to execute external decompressing programs,
387 option is specified and the built-in decompressors are not available.
388 On systems where sandboxing is not available, this option has no effect.
389 .It Fl v , Fl Fl version
390 Print the version of the program and exit.
391 .It Fl z , Fl Fl uncompress
392 Try to look inside compressed files.
393 .It Fl Z , Fl Fl uncompress-noreport
394 Try to look inside compressed files, but report information about the contents
395 only not the compression.
396 .It Fl 0 , Fl Fl print0
397 Output a null character
399 after the end of the filename.
403 This does not affect the separator, which is still printed.
405 If this option is repeated more than once, then
407 prints just the filename followed by a NUL followed by the description
408 (or ERROR: text) followed by a second NUL for each entry.
410 Print a help message and exit.
413 The environment variable
415 can be used to set the default magic file name.
416 If that variable is set, then
418 will not attempt to open
423 to the value of this variable as appropriate.
424 The environment variable
426 controls (on systems that support symbolic links), whether
428 will attempt to follow symlinks or not.
431 follows symlink, otherwise it does not.
432 This is also controlled by the
438 .Bl -tag -width __MAGIC__.mgc -compact
440 Default compiled list of magic.
442 Directory containing default magic files.
448 if the operation was successful or
450 if an error was encountered.
451 The following errors cause diagnostic messages, but don't affect the program
452 exit code (as POSIX requires), unless
455 .Bl -bullet -compact -offset indent
457 A file cannot be found
459 There is no permission to read a file
461 The file type cannot be determined
464 .Bd -literal -offset indent
465 $ file file.c file /dev/{wd0a,hda}
466 file.c: C program text
467 file: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV),
468 dynamically linked (uses shared libs), stripped
469 /dev/wd0a: block special (0/0)
470 /dev/hda: block special (3/0)
472 $ file -s /dev/wd0{b,d}
474 /dev/wd0d: x86 boot sector
476 $ file -s /dev/hda{,1,2,3,4,5,6,7,8,9,10}
477 /dev/hda: x86 boot sector
478 /dev/hda1: Linux/i386 ext2 filesystem
479 /dev/hda2: x86 boot sector
480 /dev/hda3: x86 boot sector, extended partition table
481 /dev/hda4: Linux/i386 ext2 filesystem
482 /dev/hda5: Linux/i386 swap file
483 /dev/hda6: Linux/i386 swap file
484 /dev/hda7: Linux/i386 swap file
485 /dev/hda8: Linux/i386 swap file
489 $ file -i file.c file /dev/{wd0a,hda}
491 file: application/x-executable
492 /dev/hda: application/x-not-regular-file
493 /dev/wd0a: application/x-not-regular-file
500 .Xr magic __FSECTION__
501 .Sh STANDARDS CONFORMANCE
502 This program is believed to exceed the System V Interface Definition
503 of FILE(CMD), as near as one can determine from the vague language
505 Its behavior is mostly compatible with the System V program of the same name.
506 This version knows more magic, however, so it will produce
507 different (albeit more accurate) output in many cases.
508 .\" URL: http://www.opengroup.org/onlinepubs/009695399/utilities/file.html
510 The one significant difference
511 between this version and System V
512 is that this version treats any white space
513 as a delimiter, so that spaces in pattern strings must be escaped.
515 .Bd -literal -offset indent
516 \*[Gt]10 string language impress\ (imPRESS data)
519 in an existing magic file would have to be changed to
520 .Bd -literal -offset indent
521 \*[Gt]10 string language\e impress (imPRESS data)
524 In addition, in this version, if a pattern string contains a backslash,
527 .Bd -literal -offset indent
528 0 string \ebegindata Andrew Toolkit document
531 in an existing magic file would have to be changed to
532 .Bd -literal -offset indent
533 0 string \e\ebegindata Andrew Toolkit document
536 SunOS releases 3.2 and later from Sun Microsystems include a
538 command derived from the System V one, but with some extensions.
539 This version differs from Sun's only in minor ways.
540 It includes the extension of the
544 .Bd -literal -offset indent
545 \*[Gt]16 long\*[Am]0x7fffffff \*[Gt]0 not stripped
548 On systems where libseccomp
549 .Pa ( https://github.com/seccomp/libseccomp )
552 is enforces limiting system calls to only the ones necessary for the
553 operation of the program.
554 This enforcement does not provide any security benefit when
556 is asked to decompress input files running external programs with
560 To enable execution of external decompressors, one needs to disable
565 The magic file entries have been collected from various sources,
566 mainly USENET, and contributed by various authors.
567 Christos Zoulas (address below) will collect additional
568 or corrected magic file entries.
569 A consolidation of magic file entries
570 will be distributed periodically.
572 The order of entries in the magic file is significant.
573 Depending on what system you are using, the order that
574 they are put together may be incorrect.
577 command uses a magic file,
578 keep the old magic file around for comparison purposes
580 .Pa __MAGIC__.orig ) .
585 .Dv UNIX since at least Research Version 4
586 (man page dated November, 1973).
587 The System V version introduced one significant major change:
588 the external list of magic types.
589 This slowed the program down slightly but made it a lot more flexible.
591 This program, based on the System V version,
592 was written by Ian Darwin
593 .Aq ian@darwinsys.com
594 without looking at anybody else's source code.
596 John Gilmore revised the code extensively, making it better than
598 Geoff Collyer found several inadequacies
599 and provided some magic file entries.
602 operator by Rob McMahon,
603 .Aq cudcv@warwick.ac.uk ,
608 made many changes from 1993 to the present.
610 Primary development and maintenance from 1990 to the present by
612 .Aq christos@astron.com .
614 Altered by Chris Lowth
615 .Aq chris@lowth.com ,
618 option to output mime type strings, using an alternative
619 magic file and internal logic.
621 Altered by Eric Fischer
624 to identify character codes and attempt to identify the languages
627 Altered by Reuben Thomas
629 2007-2011, to improve MIME support, merge MIME and non-MIME magic,
630 support directories as well as files of magic, apply many bug fixes,
631 update and fix a lot of magic, improve the build system, improve the
632 documentation, and rewrite the Python bindings in pure Python.
634 The list of contributors to the
636 directory (magic files)
637 is too long to include here.
638 You know who you are; thank you.
639 Many contributors are listed in the source files.
641 Copyright (c) Ian F. Darwin, Toronto, Canada, 1986-1999.
642 Covered by the standard Berkeley Software Distribution copyright; see the file
643 COPYING in the source distribution.
649 were written by John Gilmore from his public-domain
651 program, and are not covered by the above license.
653 Please report bugs and send patches to the bug tracker at
654 .Pa https://bugs.astron.com/
655 or the mailing list at
658 .Pa https://mailman.astron.com/mailman/listinfo/file
661 Fix output so that tests for MIME and APPLE flags are not needed all
662 over the place, and actual output is only done in one place.
664 Suggestion: push possible outputs on to a list, then pick the
665 last-pushed (most specific, one hopes) value at the end, or
666 use a default if the list is empty.
667 This should not slow down evaluation.
671 and printing \e012- between entries is clumsy and complicated; refactor
674 Some of the encoding logic is hard-coded in encoding.c and can be moved
675 to the magic files if we had a !:charset annotation.
677 Continue to squash all magic bugs.
678 See Debian BTS for a good source.
680 Store arbitrarily long strings, for example for %s patterns, so that
681 they can be printed out.
682 Fixes Debian bug #271672.
683 This can be done by allocating strings in a string pool, storing the
684 string pool at the end of the magic file and converting all the string
685 pointers to relative offsets from the string pool.
687 Add syntax for relative offsets after current level (Debian bug #466037).
689 Make file -ki work, i.e. give multiple MIME types.
691 Add a zip library so we can peek inside Office2007 documents to
692 print more details about their contents.
694 Add an option to print URLs for the sources of the file descriptions.
696 Combine script searches and add a way to map executable names to MIME
697 types (e.g. have a magic value for !:mime which causes the resulting
698 string to be looked up in a table).
699 This would avoid adding the same magic repeatedly for each new
700 hash-bang interpreter.
702 When a file descriptor is available, we can skip and adjust the buffer
703 instead of the hacky buffer management we do now.
709 to check for consistency at compile time (duplicate
712 pointing to undefined
719 more efficient by keeping a sorted list of names.
720 Special-case ^ to flip endianness in the parser so that it does not
721 have to be escaped, and document it.
723 If the offsets specified internally in the file exceed the buffer size
726 variable in file.h), then we don't seek to that offset, but we give up.
727 It would be better if buffer managements was done when the file descriptor
728 is available so we can seek around the file.
729 One must be careful though because this has performance and thus security
730 considerations, because one can slow down things by repeateadly seeking.
732 There is support now for keeping separate buffers and having offsets from
733 the end of the file, but the internal buffer management still needs an
736 You can obtain the original author's latest version by anonymous FTP
740 .Pa /pub/file/file-X.YZ.tar.gz .