xref: /freebsd/contrib/file/doc/magic.man (revision f976241773df2260e6170317080761d1c5814fe5)
1.\" $File: magic.man,v 1.97 2019/11/15 21:03:14 christos Exp $
2.Dd January 21, 2019
3.Dt MAGIC __FSECTION__
4.Os
5.\" install as magic.4 on USG, magic.5 on V7, Berkeley and Linux systems.
6.Sh NAME
7.Nm magic
8.Nd file command's magic pattern file
9.Sh DESCRIPTION
10This manual page documents the format of magic files as
11used by the
12.Xr file __CSECTION__
13command, version __VERSION__.
14The
15.Xr file __CSECTION__
16command identifies the type of a file using,
17among other tests,
18a test for whether the file contains certain
19.Dq "magic patterns" .
20The database of these
21.Dq "magic patterns"
22is usually located in a binary file in
23.Pa __MAGIC__.mgc
24or a directory of source text magic pattern fragment files in
25.Pa __MAGIC__ .
26The database specifies what patterns are to be tested for, what message or
27MIME type to print if a particular pattern is found,
28and additional information to extract from the file.
29.Pp
30The format of the source fragment files that are used to build this database
31is as follows:
32Each line of a fragment file specifies a test to be performed.
33A test compares the data starting at a particular offset
34in the file with a byte value, a string or a numeric value.
35If the test succeeds, a message is printed.
36The line consists of the following fields:
37.Bl -tag -width ".Dv message"
38.It Dv offset
39A number specifying the offset (in bytes) into the file of the data
40which is to be tested.
41This offset can be a negative number if it is:
42.Bl -bullet  -compact
43.It
44The first direct offset of the magic entry (at continuation level 0),
45in which case it is interpreted an offset from end end of the file
46going backwards.
47This works only when a file descriptor to the file is available and it
48is a regular file.
49.It
50A continuation offset relative to the end of the last up-level field
51.Dv ( \*[Am] ) .
52.El
53.It Dv type
54The type of the data to be tested.
55The possible values are:
56.Bl -tag -width ".Dv lestring16"
57.It Dv byte
58A one-byte value.
59.It Dv short
60A two-byte value in this machine's native byte order.
61.It Dv long
62A four-byte value in this machine's native byte order.
63.It Dv quad
64An eight-byte value in this machine's native byte order.
65.It Dv float
66A 32-bit single precision IEEE floating point number in this machine's native byte order.
67.It Dv double
68A 64-bit double precision IEEE floating point number in this machine's native byte order.
69.It Dv string
70A string of bytes.
71The string type specification can be optionally followed
72by /[WwcCtbT]*.
73The
74.Dq W
75flag compacts whitespace in the target, which must
76contain at least one whitespace character.
77If the magic has
78.Dv n
79consecutive blanks, the target needs at least
80.Dv n
81consecutive blanks to match.
82The
83.Dq w
84flag treats every blank in the magic as an optional blank.
85The
86.Dq c
87flag specifies case insensitive matching: lower case
88characters in the magic match both lower and upper case characters in the
89target, whereas upper case characters in the magic only match upper case
90characters in the target.
91The
92.Dq C
93flag specifies case insensitive matching: upper case
94characters in the magic match both lower and upper case characters in the
95target, whereas lower case characters in the magic only match upper case
96characters in the target.
97To do a complete case insensitive match, specify both
98.Dq c
99and
100.Dq C .
101The
102.Dq t
103flag forces the test to be done for text files, while the
104.Dq b
105flag forces the test to be done for binary files.
106The
107.Dq T
108flag causes the string to be trimmed, i.e. leading and trailing whitespace
109is deleted before the string is printed.
110.It Dv pstring
111A Pascal-style string where the first byte/short/int is interpreted as the
112unsigned length.
113The length defaults to byte and can be specified as a modifier.
114The following modifiers are supported:
115.Bl -tag -compact -width B
116.It B
117A byte length (default).
118.It H
119A 2 byte big endian length.
120.It h
121A 2 byte little endian length.
122.It L
123A 4 byte big endian length.
124.It l
125A 4 byte little endian length.
126.It J
127The length includes itself in its count.
128.El
129The string is not NUL terminated.
130.Dq J
131is used rather than the more
132valuable
133.Dq I
134because this type of length is a feature of the JPEG
135format.
136.It Dv date
137A four-byte value interpreted as a UNIX date.
138.It Dv qdate
139An eight-byte value interpreted as a UNIX date.
140.It Dv ldate
141A four-byte value interpreted as a UNIX-style date, but interpreted as
142local time rather than UTC.
143.It Dv qldate
144An eight-byte value interpreted as a UNIX-style date, but interpreted as
145local time rather than UTC.
146.It Dv qwdate
147An eight-byte value interpreted as a Windows-style date.
148.It Dv beid3
149A 32-bit ID3 length in big-endian byte order.
150.It Dv beshort
151A two-byte value in big-endian byte order.
152.It Dv belong
153A four-byte value in big-endian byte order.
154.It Dv bequad
155An eight-byte value in big-endian byte order.
156.It Dv befloat
157A 32-bit single precision IEEE floating point number in big-endian byte order.
158.It Dv bedouble
159A 64-bit double precision IEEE floating point number in big-endian byte order.
160.It Dv bedate
161A four-byte value in big-endian byte order,
162interpreted as a Unix date.
163.It Dv beqdate
164An eight-byte value in big-endian byte order,
165interpreted as a Unix date.
166.It Dv beldate
167A four-byte value in big-endian byte order,
168interpreted as a UNIX-style date, but interpreted as local time rather
169than UTC.
170.It Dv beqldate
171An eight-byte value in big-endian byte order,
172interpreted as a UNIX-style date, but interpreted as local time rather
173than UTC.
174.It Dv beqwdate
175An eight-byte value in big-endian byte order,
176interpreted as a Windows-style date.
177.It Dv bestring16
178A two-byte unicode (UCS16) string in big-endian byte order.
179.It Dv leid3
180A 32-bit ID3 length in little-endian byte order.
181.It Dv leshort
182A two-byte value in little-endian byte order.
183.It Dv lelong
184A four-byte value in little-endian byte order.
185.It Dv lequad
186An eight-byte value in little-endian byte order.
187.It Dv lefloat
188A 32-bit single precision IEEE floating point number in little-endian byte order.
189.It Dv ledouble
190A 64-bit double precision IEEE floating point number in little-endian byte order.
191.It Dv ledate
192A four-byte value in little-endian byte order,
193interpreted as a UNIX date.
194.It Dv leqdate
195An eight-byte value in little-endian byte order,
196interpreted as a UNIX date.
197.It Dv leldate
198A four-byte value in little-endian byte order,
199interpreted as a UNIX-style date, but interpreted as local time rather
200than UTC.
201.It Dv leqldate
202An eight-byte value in little-endian byte order,
203interpreted as a UNIX-style date, but interpreted as local time rather
204than UTC.
205.It Dv leqwdate
206An eight-byte value in little-endian byte order,
207interpreted as a Windows-style date.
208.It Dv lestring16
209A two-byte unicode (UCS16) string in little-endian byte order.
210.It Dv melong
211A four-byte value in middle-endian (PDP-11) byte order.
212.It Dv medate
213A four-byte value in middle-endian (PDP-11) byte order,
214interpreted as a UNIX date.
215.It Dv meldate
216A four-byte value in middle-endian (PDP-11) byte order,
217interpreted as a UNIX-style date, but interpreted as local time rather
218than UTC.
219.It Dv indirect
220Starting at the given offset, consult the magic database again.
221The offset of the
222.Dv indirect
223magic is by default absolute in the file, but one can specify
224.Dv /r
225to indicate that the offset is relative from the beginning of the entry.
226.It Dv name
227Define a
228.Dq named
229magic instance that can be called from another
230.Dv use
231magic entry, like a subroutine call.
232Named instance direct magic offsets are relative to the offset of the
233previous matched entry, but indirect offsets are relative to the beginning
234of the file as usual.
235Named magic entries always match.
236.It Dv use
237Recursively call the named magic starting from the current offset.
238If the name of the referenced begins with a
239.Dv ^
240then the endianness of the magic is switched; if the magic mentioned
241.Dv leshort
242for example,
243it is treated as
244.Dv beshort
245and vice versa.
246This is useful to avoid duplicating the rules for different endianness.
247.It Dv regex
248A regular expression match in extended POSIX regular expression syntax
249(like egrep).
250Regular expressions can take exponential time to process, and their
251performance is hard to predict, so their use is discouraged.
252When used in production environments, their performance
253should be carefully checked.
254The size of the string to search should also be limited by specifying
255.Dv /<length> ,
256to avoid performance issues scanning long files.
257The type specification can also be optionally followed by
258.Dv /[c][s][l] .
259The
260.Dq c
261flag makes the match case insensitive, while the
262.Dq s
263flag update the offset to the start offset of the match, rather than the end.
264The
265.Dq l
266modifier, changes the limit of length to mean number of lines instead of a
267byte count.
268Lines are delimited by the platforms native line delimiter.
269When a line count is specified, an implicit byte count also computed assuming
270each line is 80 characters long.
271If neither a byte or line count is specified, the search is limited automatically
272to 8KiB.
273.Dv ^
274and
275.Dv $
276match the beginning and end of individual lines, respectively,
277not beginning and end of file.
278.It Dv search
279A literal string search starting at the given offset.
280The same modifier flags can be used as for string patterns.
281The search expression must contain the range in the form
282.Dv /number,
283that is the number of positions at which the match will be
284attempted, starting from the start offset.
285This is suitable for
286searching larger binary expressions with variable offsets, using
287.Dv \e
288escapes for special characters.
289The order of modifier and number is not relevant.
290.It Dv default
291This is intended to be used with the test
292.Em x
293(which is always true) and it has no type.
294It matches when no other test at that continuation level has matched before.
295Clearing that matched tests for a continuation level, can be done using the
296.Dv clear
297test.
298.It Dv clear
299This test is always true and clears the match flag for that continuation level.
300It is intended to be used with the
301.Dv default
302test.
303.El
304.Pp
305For compatibility with the Single
306.Ux
307Standard, the type specifiers
308.Dv dC
309and
310.Dv d1
311are equivalent to
312.Dv byte ,
313the type specifiers
314.Dv uC
315and
316.Dv u1
317are equivalent to
318.Dv ubyte ,
319the type specifiers
320.Dv dS
321and
322.Dv d2
323are equivalent to
324.Dv short ,
325the type specifiers
326.Dv uS
327and
328.Dv u2
329are equivalent to
330.Dv ushort ,
331the type specifiers
332.Dv dI ,
333.Dv dL ,
334and
335.Dv d4
336are equivalent to
337.Dv long ,
338the type specifiers
339.Dv uI ,
340.Dv uL ,
341and
342.Dv u4
343are equivalent to
344.Dv ulong ,
345the type specifier
346.Dv d8
347is equivalent to
348.Dv quad ,
349the type specifier
350.Dv u8
351is equivalent to
352.Dv uquad ,
353and the type specifier
354.Dv s
355is equivalent to
356.Dv string .
357In addition, the type specifier
358.Dv dQ
359is equivalent to
360.Dv quad
361and the type specifier
362.Dv uQ
363is equivalent to
364.Dv uquad .
365.Pp
366Each top-level magic pattern (see below for an explanation of levels)
367is classified as text or binary according to the types used.
368Types
369.Dq regex
370and
371.Dq search
372are classified as text tests, unless non-printable characters are used
373in the pattern.
374All other tests are classified as binary.
375A top-level
376pattern is considered to be a test text when all its patterns are text
377patterns; otherwise, it is considered to be a binary pattern.
378When
379matching a file, binary patterns are tried first; if no match is
380found, and the file looks like text, then its encoding is determined
381and the text patterns are tried.
382.Pp
383The numeric types may optionally be followed by
384.Dv \*[Am]
385and a numeric value,
386to specify that the value is to be AND'ed with the
387numeric value before any comparisons are done.
388Prepending a
389.Dv u
390to the type indicates that ordered comparisons should be unsigned.
391.It Dv test
392The value to be compared with the value from the file.
393If the type is
394numeric, this value
395is specified in C form; if it is a string, it is specified as a C string
396with the usual escapes permitted (e.g. \en for new-line).
397.Pp
398Numeric values
399may be preceded by a character indicating the operation to be performed.
400It may be
401.Dv = ,
402to specify that the value from the file must equal the specified value,
403.Dv \*[Lt] ,
404to specify that the value from the file must be less than the specified
405value,
406.Dv \*[Gt] ,
407to specify that the value from the file must be greater than the specified
408value,
409.Dv \*[Am] ,
410to specify that the value from the file must have set all of the bits
411that are set in the specified value,
412.Dv ^ ,
413to specify that the value from the file must have clear any of the bits
414that are set in the specified value, or
415.Dv ~ ,
416the value specified after is negated before tested.
417.Dv x ,
418to specify that any value will match.
419If the character is omitted, it is assumed to be
420.Dv = .
421Operators
422.Dv \*[Am] ,
423.Dv ^ ,
424and
425.Dv ~
426don't work with floats and doubles.
427The operator
428.Dv !\&
429specifies that the line matches if the test does
430.Em not
431succeed.
432.Pp
433Numeric values are specified in C form; e.g.
434.Dv 13
435is decimal,
436.Dv 013
437is octal, and
438.Dv 0x13
439is hexadecimal.
440.Pp
441Numeric operations are not performed on date types, instead the numeric
442value is interpreted as an offset.
443.Pp
444For string values, the string from the
445file must match the specified string.
446The operators
447.Dv = ,
448.Dv \*[Lt]
449and
450.Dv \*[Gt]
451(but not
452.Dv \*[Am] )
453can be applied to strings.
454The length used for matching is that of the string argument
455in the magic file.
456This means that a line can match any non-empty string (usually used to
457then print the string), with
458.Em \*[Gt]\e0
459(because all non-empty strings are greater than the empty string).
460.Pp
461Dates are treated as numerical values in the respective internal
462representation.
463.Pp
464The special test
465.Em x
466always evaluates to true.
467.It Dv message
468The message to be printed if the comparison succeeds.
469If the string contains a
470.Xr printf 3
471format specification, the value from the file (with any specified masking
472performed) is printed using the message as the format string.
473If the string begins with
474.Dq \eb ,
475the message printed is the remainder of the string with no whitespace
476added before it: multiple matches are normally separated by a single
477space.
478.El
479.Pp
480An APPLE 4+4 character APPLE creator and type can be specified as:
481.Bd -literal -offset indent
482!:apple	CREATYPE
483.Ed
484.Pp
485A MIME type is given on a separate line, which must be the next
486non-blank or comment line after the magic line that identifies the
487file type, and has the following format:
488.Bd -literal -offset indent
489!:mime	MIMETYPE
490.Ed
491.Pp
492i.e. the literal string
493.Dq !:mime
494followed by the MIME type.
495.Pp
496An optional strength can be supplied on a separate line which refers to
497the current magic description using the following format:
498.Bd -literal -offset indent
499!:strength OP VALUE
500.Ed
501.Pp
502The operand
503.Dv OP
504can be:
505.Dv + ,
506.Dv - ,
507.Dv * ,
508or
509.Dv /
510and
511.Dv VALUE
512is a constant between 0 and 255.
513This constant is applied using the specified operand
514to the currently computed default magic strength.
515.Pp
516Some file formats contain additional information which is to be printed
517along with the file type or need additional tests to determine the true
518file type.
519These additional tests are introduced by one or more
520.Em \*[Gt]
521characters preceding the offset.
522The number of
523.Em \*[Gt]
524on the line indicates the level of the test; a line with no
525.Em \*[Gt]
526at the beginning is considered to be at level 0.
527Tests are arranged in a tree-like hierarchy:
528if the test on a line at level
529.Em n
530succeeds, all following tests at level
531.Em n+1
532are performed, and the messages printed if the tests succeed, until a line
533with level
534.Em n
535(or less) appears.
536For more complex files, one can use empty messages to get just the
537"if/then" effect, in the following way:
538.Bd -literal -offset indent
5390      string   MZ
540\*[Gt]0x18  leshort  \*[Lt]0x40   MS-DOS executable
541\*[Gt]0x18  leshort  \*[Gt]0x3f   extended PC executable (e.g., MS Windows)
542.Ed
543.Pp
544Offsets do not need to be constant, but can also be read from the file
545being examined.
546If the first character following the last
547.Em \*[Gt]
548is a
549.Em \&(
550then the string after the parenthesis is interpreted as an indirect offset.
551That means that the number after the parenthesis is used as an offset in
552the file.
553The value at that offset is read, and is used again as an offset
554in the file.
555Indirect offsets are of the form:
556.Em (( x [[.,][bBcCeEfFgGhHiIlmsSqQ]][+\-][ y ]) .
557The value of
558.Em x
559is used as an offset in the file.
560A byte, id3 length, short or long is read at that offset depending on the
561.Em [bBcCeEfFgGhHiIlmsSqQ]
562type specifier.
563The value is treated as signed if
564.Dq ,
565is specified or unsigned if
566.Dq .
567is specified.
568The capitalized types interpret the number as a big endian
569value, whereas the small letter versions interpret the number as a little
570endian value;
571the
572.Em m
573type interprets the number as a middle endian (PDP-11) value.
574To that number the value of
575.Em y
576is added and the result is used as an offset in the file.
577The default type if one is not specified is long.
578The following types are recognized:
579.Bl -column -offset indent "Type" "Half/Short" "Little" "Size"
580.It Sy Type	Sy Mnemonic	Sy Endian	Sy Size
581.It bcBc	Byte/Char	N/A	1
582.It efg	Double	Little	8
583.It EFG	Double	Big	8
584.It hs	Half/Short	Little	2
585.It HS	Half/Short	Big	2
586.It i	ID3	Little	4
587.It I	ID3	Big	4
588.It m	Middle	Middle	4
589.It q	Quad	Little	8
590.It Q	Quad	Big	8
591.El
592.Pp
593That way variable length structures can be examined:
594.Bd -literal -offset indent
595# MS Windows executables are also valid MS-DOS executables
5960           string  MZ
597\*[Gt]0x18       leshort \*[Lt]0x40   MZ executable (MS-DOS)
598# skip the whole block below if it is not an extended executable
599\*[Gt]0x18       leshort \*[Gt]0x3f
600\*[Gt]\*[Gt](0x3c.l)  string  PE\e0\e0  PE executable (MS-Windows)
601\*[Gt]\*[Gt](0x3c.l)  string  LX\e0\e0  LX executable (OS/2)
602.Ed
603.Pp
604This strategy of examining has a drawback: you must make sure that you
605eventually print something, or users may get empty output (such as when
606there is neither PE\e0\e0 nor LE\e0\e0 in the above example).
607.Pp
608If this indirect offset cannot be used directly, simple calculations are
609possible: appending
610.Em [+-*/%\*[Am]|^]number
611inside parentheses allows one to modify
612the value read from the file before it is used as an offset:
613.Bd -literal -offset indent
614# MS Windows executables are also valid MS-DOS executables
6150           string  MZ
616# sometimes, the value at 0x18 is less that 0x40 but there's still an
617# extended executable, simply appended to the file
618\*[Gt]0x18       leshort \*[Lt]0x40
619\*[Gt]\*[Gt](4.s*512) leshort 0x014c  COFF executable (MS-DOS, DJGPP)
620\*[Gt]\*[Gt](4.s*512) leshort !0x014c MZ executable (MS-DOS)
621.Ed
622.Pp
623Sometimes you do not know the exact offset as this depends on the length or
624position (when indirection was used before) of preceding fields.
625You can specify an offset relative to the end of the last up-level
626field using
627.Sq \*[Am]
628as a prefix to the offset:
629.Bd -literal -offset indent
6300           string  MZ
631\*[Gt]0x18       leshort \*[Gt]0x3f
632\*[Gt]\*[Gt](0x3c.l)  string  PE\e0\e0    PE executable (MS-Windows)
633# immediately following the PE signature is the CPU type
634\*[Gt]\*[Gt]\*[Gt]\*[Am]0       leshort 0x14c     for Intel 80386
635\*[Gt]\*[Gt]\*[Gt]\*[Am]0       leshort 0x184     for DEC Alpha
636.Ed
637.Pp
638Indirect and relative offsets can be combined:
639.Bd -literal -offset indent
6400             string  MZ
641\*[Gt]0x18         leshort \*[Lt]0x40
642\*[Gt]\*[Gt](4.s*512)   leshort !0x014c MZ executable (MS-DOS)
643# if it's not COFF, go back 512 bytes and add the offset taken
644# from byte 2/3, which is yet another way of finding the start
645# of the extended executable
646\*[Gt]\*[Gt]\*[Gt]\*[Am](2.s-514) string  LE      LE executable (MS Windows VxD driver)
647.Ed
648.Pp
649Or the other way around:
650.Bd -literal -offset indent
6510                 string  MZ
652\*[Gt]0x18             leshort \*[Gt]0x3f
653\*[Gt]\*[Gt](0x3c.l)        string  LE\e0\e0  LE executable (MS-Windows)
654# at offset 0x80 (-4, since relative offsets start at the end
655# of the up-level match) inside the LE header, we find the absolute
656# offset to the code area, where we look for a specific signature
657\*[Gt]\*[Gt]\*[Gt](\*[Am]0x7c.l+0x26) string  UPX     \eb, UPX compressed
658.Ed
659.Pp
660Or even both!
661.Bd -literal -offset indent
6620                string  MZ
663\*[Gt]0x18            leshort \*[Gt]0x3f
664\*[Gt]\*[Gt](0x3c.l)       string  LE\e0\e0 LE executable (MS-Windows)
665# at offset 0x58 inside the LE header, we find the relative offset
666# to a data area where we look for a specific signature
667\*[Gt]\*[Gt]\*[Gt]\*[Am](\*[Am]0x54.l-3)  string  UNACE  \eb, ACE self-extracting archive
668.Ed
669.Pp
670If you have to deal with offset/length pairs in your file, even the
671second value in a parenthesized expression can be taken from the file itself,
672using another set of parentheses.
673Note that this additional indirect offset is always relative to the
674start of the main indirect offset.
675.Bd -literal -offset indent
6760                 string       MZ
677\*[Gt]0x18             leshort      \*[Gt]0x3f
678\*[Gt]\*[Gt](0x3c.l)        string       PE\e0\e0 PE executable (MS-Windows)
679# search for the PE section called ".idata"...
680\*[Gt]\*[Gt]\*[Gt]\*[Am]0xf4          search/0x140 .idata
681# ...and go to the end of it, calculated from start+length;
682# these are located 14 and 10 bytes after the section name
683\*[Gt]\*[Gt]\*[Gt]\*[Gt](\*[Am]0xe.l+(-4)) string       PK\e3\e4 \eb, ZIP self-extracting archive
684.Ed
685.Pp
686If you have a list of known values at a particular continuation level,
687and you want to provide a switch-like default case:
688.Bd -literal -offset indent
689# clear that continuation level match
690\*[Gt]18	clear
691\*[Gt]18	lelong	1	one
692\*[Gt]18	lelong	2	two
693\*[Gt]18	default	x
694# print default match
695\*[Gt]\*[Gt]18	lelong	x	unmatched 0x%x
696.Ed
697.Sh SEE ALSO
698.Xr file __CSECTION__
699\- the command that reads this file.
700.Sh BUGS
701The formats
702.Dv long ,
703.Dv belong ,
704.Dv lelong ,
705.Dv melong ,
706.Dv short ,
707.Dv beshort ,
708and
709.Dv leshort
710do not depend on the length of the C data types
711.Dv short
712and
713.Dv long
714on the platform, even though the Single
715.Ux
716Specification implies that they do.  However, as OS X Mountain Lion has
717passed the Single
718.Ux
719Specification validation suite, and supplies a version of
720.Xr file __CSECTION__
721in which they do not depend on the sizes of the C data types and that is
722built for a 64-bit environment in which
723.Dv long
724is 8 bytes rather than 4 bytes, presumably the validation suite does not
725test whether, for example
726.Dv long
727refers to an item with the same size as the C data type
728.Dv long .
729There should probably be
730.Dv type
731names
732.Dv int8 ,
733.Dv uint8 ,
734.Dv int16 ,
735.Dv uint16 ,
736.Dv int32 ,
737.Dv uint32 ,
738.Dv int64 ,
739and
740.Dv uint64 ,
741and specified-byte-order variants of them,
742to make it clearer that those types have specified widths.
743.\"
744.\" From: guy@sun.uucp (Guy Harris)
745.\" Newsgroups: net.bugs.usg
746.\" Subject: /etc/magic's format isn't well documented
747.\" Message-ID: <2752@sun.uucp>
748.\" Date: 3 Sep 85 08:19:07 GMT
749.\" Organization: Sun Microsystems, Inc.
750.\" Lines: 136
751.\"
752.\" Here's a manual page for the format accepted by the "file" made by adding
753.\" the changes I posted to the S5R2 version.
754.\"
755.\" Modified for Ian Darwin's version of the file command.
756