Index: head/usr.bin/file/file.1
===================================================================
--- head/usr.bin/file/file.1	(revision 169976)
+++ head/usr.bin/file/file.1	(revision 169977)
@@ -1,588 +1,588 @@
 .\" $FreeBSD$
 .\" $Id: file.man,v 1.57 2005/08/18 15:18:22 christos Exp $
 .Dd August 18, 2005
 .Dt FILE 1 "Copyright but distributable"
 .Os
 .Sh NAME
 .Nm file
 .Nd determine file type
 .Sh SYNOPSIS
 .Nm
 .Op Fl bchikLnNprsvz
 .Op Fl f Ar namefile
 .Op Fl F Ar separator
 .Op Fl m Ar magicfiles
 .Ar
 .Nm
 .Fl C
 .Op Fl m Ar magicfile
 .Sh DESCRIPTION
-This manual page documents version 4.17 of the
+This manual page documents version 4.21 of the
 .Nm
 utility which tests each argument in an attempt to classify it.
 There are three sets of tests, performed in this order:
 file system tests, magic number tests, and language tests.
 The
 .Em first
 test that succeeds causes the file type to be printed.
 .Pp
 The type printed will usually contain one of the words
 .Dq Li text
 (the file contains only
 printing characters and a few common control
 characters and is probably safe to read on an
 .Tn ASCII
 terminal),
 .Dq Li executable
 (the file contains the result of compiling a program
 in a form understandable to some
 .Ux
 kernel or another),
 or
 .Dq Li data
 meaning anything else (data is usually
 .Sq binary
 or non-printable).
 Exceptions are well-known file formats (core files, tar archives)
 that are known to contain binary data.
 When modifying the file
 .Pa /usr/share/misc/magic
 or the program itself,
 .Em "preserve these keywords" .
 People depend on knowing that all the readable files in a directory
 have the word
 .Dq Li text
 printed.
 Do not do as Berkeley did and change
 .Dq Li "shell commands text"
 to
 .Dq Li "shell script" .
 Note that the file
 .Pa /usr/share/misc/magic
 is built mechanically from a large number of small files in
 the subdirectory
 .Pa Magdir
 in the source distribution of this program.
 .Pp
 The file system tests are based on examining the return from a
 .Xr stat 2
 system call.
 The program checks to see if the file is empty,
 or if it is some sort of special file.
 Any known file types appropriate to the system you are running on
 (sockets, symbolic links, or named pipes (FIFOs) on those systems that
 implement them)
 are intuited if they are defined in
 the system header file
 .In sys/stat.h .
 .Pp
 The magic number tests are used to check for files with data in
 particular fixed formats.
 The canonical example of this is a binary executable (compiled program)
 .Pa a.out
 file, whose format is defined in
 .In a.out.h
 and possibly
 .In exec.h
 in the standard include directory.
 These files have a
 .Sq "magic number"
 stored in a particular place
 near the beginning of the file that tells the
 .Ux
 operating system
 that the file is a binary executable, and which of several types thereof.
 The concept of
 .Sq "magic number"
 has been applied by extension to data files.
 Any file with some invariant identifier at a small fixed
 offset into the file can usually be described in this way.
 The information identifying these files is read from the compiled
 magic file
 .Pa /usr/share/misc/magic.mgc ,
 or
 .Pa /usr/share/misc/magic
 if the compile file does not exist.
 In addition
 .Nm
 will look in
 .Pa $HOME/.magic.mgc ,
 or
 .Pa $HOME/.magic
 for magic entries.
 .Pp
 If a file does not match any of the entries in the magic file,
 it is examined to see if it seems to be a text file.
 ASCII, ISO-8859-x, non-ISO 8-bit extended-ASCII character sets
 (such as those used on Macintosh and IBM PC systems),
 UTF-8-encoded Unicode, UTF-16-encoded Unicode, and EBCDIC
 character sets can be distinguished by the different
 ranges and sequences of bytes that constitute printable text
 in each set.
 If a file passes any of these tests, its character set is reported.
 ASCII, ISO-8859-x, UTF-8, and extended-ASCII files are identified
 as
 .Dq Li text
 because they will be mostly readable on nearly any terminal;
 UTF-16 and EBCDIC are only
 .Dq Li "character data"
 because, while
 they contain text, it is text that will require translation
 before it can be read.
 In addition,
 .Nm
 will attempt to determine other characteristics of text-type files.
 If the lines of a file are terminated by CR, CRLF, or NEL, instead
 of the
 .Ux Ns -standard
 LF, this will be reported.
 Files that contain embedded escape sequences or overstriking
 will also be identified.
 .Pp
 Once
 .Nm
 has determined the character set used in a text-type file,
 it will
 attempt to determine in what language the file is written.
 The language tests look for particular strings (cf
 .Pa names.h )
 that can appear anywhere in the first few blocks of a file.
 For example, the keyword
 .Ic .br
 indicates that the file is most likely a
 .Xr troff 1
 input file, just as the keyword
 .Ic struct
 indicates a C program.
 These tests are less reliable than the previous
 two groups, so they are performed last.
 The language test routines also test for some miscellany
 (such as
 .Xr tar 1
 archives).
 .Pp
 Any file that cannot be identified as having been written
 in any of the character sets listed above is simply said to be
 .Dq Li data .
 .Sh OPTIONS
 .Bl -tag -width indent
 .It Fl b , -brief
 Do not prepend filenames to output lines (brief mode).
 .It Fl c , -checking-printout
 Cause a checking printout of the parsed form of the magic file.
 This is usually used in conjunction with
 .Fl m
 to debug a new magic file before installing it.
 .It Fl C , -compile
 Write a
 .Pa magic.mgc
 output file that contains a pre-parsed version of
 file.
 .It Fl f , -files-from Ar namefile
 Read the names of the files to be examined from
 .Ar namefile
 (one per line)
 before the argument list.
 Either
 .Ar namefile
 or at least one filename argument must be present;
 to test the standard input, use
 .Dq Fl
 as a filename argument.
 .It Fl F , -separator Ar separator
 Use the specified string as the separator between the filename and the
 file result returned.
 Defaults to
 .Ql \&: .
 .It Fl h , -no-dereference
 Causes symlinks not to be followed
 (on systems that support symbolic links).
 This is the default if the
 environment variable
 .Ev POSIXLY_CORRECT
 is not defined.
 .It Fl i , -mime
 Causes the file command to output mime type strings rather than the more
 traditional human readable ones.
 Thus it may say
 .Dq Li "text/plain; charset=us-ascii"
 rather than
 .Dq Li "ASCII text" .
 In order for this option to work, file changes the way
 it handles files recognised by the command itself (such as many of the
 text file types, directories etc), and makes use of an alternative
 .Pa magic
 file.
 (See
 .Sx FILES
 section, below).
 .It Fl k , -keep-going
 Do not stop at the first match, keep going.
 .It Fl L , -dereference
 option causes symlinks to be followed, as the like-named option in
 .Xr ls 1
 (on systems that support symbolic links).
 This is the default if the environment variable
 .Ev POSIXLY_CORRECT
 is defined.
 .It Fl m , -magic-file Ar list
 Specify an alternate list of files containing magic numbers.
 This can be a single file, or a colon-separated list of files.
 If a compiled magic file is found alongside, it will be used instead.
 With the
 .Fl i
 or
 .Fl -mime
 option, the program adds
 .Pa .mime
 to each file name.
 .It Fl n , -no-buffer
 Force stdout to be flushed after checking each file.
 This is only useful if checking a list of files.
 It is intended to be used by programs that want
 filetype output from a pipe.
 .It Fl N , -no-pad
 Do not pad filenames so that they align in the output.
 .It Fl p , -preserve-date
 On systems that support
 .Xr utime 3
 or
 .Xr utimes 2 ,
 attempt to preserve the access time of files analyzed, to pretend that
 .Nm
 never read them.
 .It Fl r , -raw
 Do not translate unprintable characters to \eooo.
 Normally
 .Nm
 translates unprintable characters to their octal representation.
 .It Fl s , -special-files
 Normally,
 .Nm
 only attempts to read and determine the type of argument files which
 .Xr stat 2
 reports are ordinary files.
 This prevents problems, because reading special files may have peculiar
 consequences.
 Specifying the
 .Fl s
 option causes
 .Nm
 to also read argument files which are block or character special files.
 This is useful for determining the file system types of the data in raw
 disk partitions, which are block special files.
 This option also causes
 .Nm
 to disregard the file size as reported by
 .Xr stat 2
 since on some systems it reports a zero size for raw disk partitions.
 .It Fl v , -version
 Print the version of the program and exit.
 .It Fl z , -uncompress
 Try to look inside compressed files.
 .It Fl -help
 Print a help message and exit.
 .El
 .Sh FILES
 .Bl -tag -width ".Pa /usr/share/misc/magic.mime" -compact
 .It Pa /usr/share/misc/magic.mgc
 Default compiled list of magic numbers
 .It Pa /usr/share/misc/magic
 Default list of magic numbers
 .It Pa /usr/share/misc/magic.mime.mgc
 Default compiled list of magic numbers, used to output mime types when
 the
 .Fl i
 option is specified.
 .It Pa /usr/share/misc/magic.mime
 Default list of magic numbers, used to output mime types when the
 .Fl i
 option is specified.
 .El
 .Sh ENVIRONMENT
 The environment variable
 .Ev MAGIC
 can be used to set the default magic number file name.
 If that variable is set, then
 .Nm
 will not attempt to open
 .Pa $HOME/.magic .
 .Nm
 adds
 .Pa .mime
 and/or
 .Pa .mgc
 to the value of this variable as appropriate.
 The environment variable
 .Ev POSIXLY_CORRECT
 controls (on systems that support symbolic links), if
 .Nm
 will attempt to follow symlinks or not.
 If set, then
 .Nm
 follows symlink, otherwise it does not.
 This is also controlled
 by the
 .Fl L
 and
 .Fl h
 options.
 .Sh SEE ALSO
 .Xr hexdump 1 ,
 .Xr od 1 ,
 .Xr strings 1 ,
 .Xr magic 5
 .Sh STANDARDS CONFORMANCE
 This program is believed to exceed the
 .St -svid4
 of FILE(CMD), as near as one can determine from the vague language
 contained therein.
 Its behaviour is mostly compatible with the System V program of the same name.
 This version knows more magic, however, so it will produce
 different (albeit more accurate) output in many cases.
 .Pp
 The one significant difference
 between this version and System V
 is that this version treats any white space
 as a delimiter, so that spaces in pattern strings must be escaped.
 For example,
 .Pp
 .Dl ">10	string	language impress\ 	(imPRESS data)"
 .Pp
 in an existing magic file would have to be changed to
 .Pp
 .Dl ">10	string	language\e impress	(imPRESS data)"
 .Pp
 In addition, in this version, if a pattern string contains a backslash,
 it must be escaped.
 For example
 .Pp
 .Dl "0	string		\ebegindata	Andrew Toolkit document"
 .Pp
 in an existing magic file would have to be changed to
 .Pp
 .Dl "0	string		\e\ebegindata	Andrew Toolkit document"
 .Pp
 SunOS releases 3.2 and later from Sun Microsystems include a
 .Xr file 1
 command derived from the System V one, but with some extensions.
 My version differs from Sun's only in minor ways.
 It includes the extension of the
 .Sq Ic &
 operator, used as,
 for example,
 .Pp
 .Dl ">16	long&0x7fffffff	>0		not stripped"
 .Sh MAGIC DIRECTORY
 The magic file entries have been collected from various sources,
 mainly USENET, and contributed by various authors.
 .An Christos Zoulas
 (address below) will collect additional
 or corrected magic file entries.
 A consolidation of magic file entries
 will be distributed periodically.
 .Pp
 The order of entries in the magic file is significant.
 Depending on what system you are using, the order that
 they are put together may be incorrect.
 If your old
 .Nm
 command uses a magic file,
 keep the old magic file around for comparison purposes
 (rename it to
 .Pa /usr/share/misc/magic.orig ) .
 .Sh EXAMPLES
 .Bd -literal
 $ file file.c file /dev/{wd0a,hda}
 file.c:    C program text
 file:      ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV),
            dynamically linked (uses shared libs), stripped
 /dev/wd0a: block special (0/0)
 /dev/hda:  block special (3/0)
 $ file -s /dev/wd0{b,d}
 /dev/wd0b: data
 /dev/wd0d: x86 boot sector
 $ file -s /dev/hda{,1,2,3,4,5,6,7,8,9,10}
 /dev/hda:   x86 boot sector
 /dev/hda1:  Linux/i386 ext2 filesystem
 /dev/hda2:  x86 boot sector
 /dev/hda3:  x86 boot sector, extended partition table
 /dev/hda4:  Linux/i386 ext2 filesystem
 /dev/hda5:  Linux/i386 swap file
 /dev/hda6:  Linux/i386 swap file
 /dev/hda7:  Linux/i386 swap file
 /dev/hda8:  Linux/i386 swap file
 /dev/hda9:  empty
 /dev/hda10: empty
 
 $ file -i file.c file /dev/{wd0a,hda}
 file.c:    text/x-c
 file:      application/x-executable, dynamically linked (uses shared libs),
            not stripped
 /dev/hda:  application/x-not-regular-file
 /dev/wd0a: application/x-not-regular-file
 .Ed
 .Sh HISTORY
 There has been a
 .Nm
 command in every
 .Ux
 since at least Research Version 4
 (man page dated November, 1973).
 The System V version introduced one significant major change:
 the external list of magic number types.
 This slowed the program down slightly but made it a lot more flexible.
 .Pp
 This program, based on the System V version,
 was written by
 .An Ian Darwin Aq ian@darwinsys.com
 without looking at anybody else's source code.
 .Pp
 .An John Gilmore
 revised the code extensively, making it better than
 the first version.
 .An Geoff Collyer
 found several inadequacies
 and provided some magic file entries.
 Contributions by the
 .Sq Ic &
 operator by
 .An Rob McMahon Aq cudcv@warwick.ac.uk ,
 1989.
 .Pp
 .An Guy Harris Aq guy@netapp.com ,
 made many changes from 1993 to the present.
 .Pp
 Primary development and maintenance from 1990 to the present by
 .An Christos Zoulas Aq christos@astron.com .
 .Pp
 Altered by
 .An Chris Lowth Aq chris@lowth.com ,
 2000:
 Handle the
 .Fl i
 option to output mime type strings and using an alternative
 magic file and internal logic.
 .Pp
 Altered by
 .An Eric Fischer Aq enf@pobox.com ,
 July, 2000,
 to identify character codes and attempt to identify the languages
 of
 .No non- Ns Tn ASCII
 files.
 .Pp
 The list of contributors to the
 .Pa Magdir
 directory (source for the
 .Pa /usr/share/misc/magic
 file) is too long to include here.
 You know who you are; thank you.
 .Sh LEGAL NOTICE
 Copyright (c)
 .An Ian F. Darwin ,
 Toronto, Canada, 1986-1999.
 Covered by the standard Berkeley Software Distribution copyright; see the file
 .Pa LEGAL.NOTICE
 in the source distribution.
 .Pp
 The files
 .Pa tar.h
 and
 .Pa is_tar.c
 were written by
 .An John Gilmore
 from his public-domain
 .Nm tar
 program, and are not covered by the above license.
 .Sh BUGS
 There must be a better way to automate the construction of the
 .Pa Magic
 file from all the glop in
 .Pa Magdir .
 What is it?
 Better yet, the magic file should be compiled into binary (say,
 .Xr ndbm 3
 or, better yet, fixed-length
 .Tn ASCII
 strings for use in heterogenous network environments) for faster startup.
 Then the program would run as fast as the Version 7 program of the same name,
 with the flexibility of the System V version.
 .Pp
 The
 .Nm
 utility uses several algorithms that favor speed over accuracy,
 thus it can be misled about the contents of
 text
 files.
 .Pp
 The support for
 text
 files (primarily for programming languages)
 is simplistic, inefficient and requires recompilation to update.
 .Pp
 There should be an
 .Ic else
 clause to follow a series of continuation lines.
 .Pp
 The magic file and keywords should have regular expression support.
 Their use of
 .Tn "ASCII TAB"
 as a field delimiter is ugly and makes
 it hard to edit the files, but is entrenched.
 .Pp
 It might be advisable to allow upper-case letters in keywords
 for e.g.,
 .Xr troff 1
 commands vs man page macros.
 Regular expression support would make this easy.
 .Pp
 The program does not grok
 .Tn FORTRAN .
 It should be able to figure
 .Tn FORTRAN
 by seeing some keywords which
 appear indented at the start of line.
 Regular expression support would make this easy.
 .Pp
 The list of keywords in
 .Pa ascmagic
 probably belongs in the
 .Pa Magic
 file.
 This could be done by using some keyword like
 .Sq Ic *
 for the offset value.
 .Pp
 Another optimisation would be to sort
 the magic file so that we can just run down all the
 tests for the first byte, first word, first long, etc, once we
 have fetched it.
 Complain about conflicts in the magic file entries.
 Make a rule that the magic entries sort based on file offset rather
 than position within the magic file?
 .Pp
 The program should provide a way to give an estimate
 of
 .Dq how good
 a guess is.
 We end up removing guesses (e.g.\&
 .Dq Li "From "
 as first 5 chars of file) because
 they are not as good as other guesses (e.g.\&
 .Dq Li "Newsgroups:"
 versus
 .Dq Li "Return-Path:" ) .
 Still, if the others do not pan out, it should be possible to use the
 first guess.
 .Pp
 This program is slower than some vendors' file commands.
 The new support for multiple character codes makes it even slower.
 .Pp
 This manual page, and particularly this section, is too long.
 .Sh AVAILABILITY
 You can obtain the original author's latest version by anonymous FTP
 on
 .Pa ftp.astron.com
 in the directory
 .Pa /pub/file/file-X.YZ.tar.gz
Index: head/usr.bin/file/magic.5
===================================================================
--- head/usr.bin/file/magic.5	(revision 169976)
+++ head/usr.bin/file/magic.5	(revision 169977)
@@ -1,444 +1,444 @@
 .\"
 .\" $FreeBSD$
 .\"
 .\" install as magic.4 on USG, magic.5 on V7 or Berkeley systems.
 .\"
 .Dd February 19, 2006
 .Dt MAGIC 5 "Public Domain"
 .Os
 .Sh NAME
 .Nm magic
 .Nd file command's magic number file
 .Sh DESCRIPTION
 This manual page documents the format of the magic file as
 used by the
 .Nm
-command, version 4.17.
+command, version 4.21.
 The
 .Nm file
 command identifies the type of a file using,
 among other tests,
 a test for whether the file begins with a certain
 .Em "magic number" .
 The file
 .Pa /usr/share/misc/magic
 specifies what magic numbers are to be tested for,
 what message to print if a particular magic number is found,
 and additional information to extract from the file.
 .Pp
 Each line of the file specifies a test to be performed.
 A test compares the data starting at a particular offset
 in the file with a 1-byte, 2-byte, or 4-byte numeric value or
 a string.
 If the test succeeds, a message is printed.
 The line consists of the following fields:
 .Bl -tag -width indent
 .It offset
 A number specifying the offset, in bytes, into the file of the data
 which is to be tested.
 .It type
 The type of the data to be tested.
 The possible values are:
 .Bl -tag -width indent
 .It byte
 A one-byte value.
 .It short
 A two-byte value (on most systems) in this machine's native byte order.
 .It long
 A four-byte value (on most systems) in this machine's native byte order.
 .It string
 A string of bytes.
 The string type specification can be optionally followed
 by /[Bbc]*.
 The
 .Dq B
 flag compacts whitespace in the target, which must contain
 at least one whitespace character.
 If the magic has
 .Ar n
 consecutive blanks, the target needs at least
 .Ar n
 consecutive blanks to match.
 The
 .Dq b
 flag treats every blank in the target as an optional blank.
 Finally the
 .Dq c
 flag, specifies case insensitive matching: lowercase characters
 in the magic match both lower and upper case characters in the
 targer, whereas upper case characters in the magic, only much
 uppercase characters in the target.
 .It pstring
 A pascal style string where the first byte is interpreted as the an
 unsigned length.
 The string is not
 .Dv NUL
 terminated.
 .It date
 A four-byte value interpreted as a
 .Ux
 date.
 .It ldate
 A four-byte value interpreted as a
 .Ux Ns -style
 date, but interpreted as
 local time rather than UTC.
 .It beshort
 A two-byte value (on most systems) in big-endian byte order.
 .It belong
 A four-byte value (on most systems) in big-endian byte order.
 .It bedate
 A four-byte value (on most systems) in big-endian byte order,
 interpreted as a
 .Ux
 date.
 .It beldate
 A four-byte value (on most systems) in big-endian byte order,
 interpreted as a
 .Ux Ns -style
 date, but interpreted as local time rather
 than UTC.
 .It bestring16
 A two-byte unicode (UCS16) string in big-endian byte order.
 .It leshort
 A two-byte value (on most systems) in little-endian byte order.
 .It lelong
 A four-byte value (on most systems) in little-endian byte order.
 .It ledate
 A four-byte value (on most systems) in little-endian byte order,
 interpreted as a
 .Ux
 date.
 .It leldate
 A four-byte value (on most systems) in little-endian byte order,
 interpreted as a
 .Ux Ns -style
 date, but interpreted as local time rather
 than UTC.
 .It lestring16
 A two-byte unicode (UCS16) string in little-endian byte order.
 .It melong
 A four-byte value (on most systems) in middle-endian (PDP-11) byte order.
 .It medate
 A four-byte value (on most systems) in middle-endian (PDP-11) byte order,
 interpreted as a
 .Ux
 date.
 .It meldate
 A four-byte value (on most systems) in middle-endian (PDP-11) byte order,
 interpreted as a
 .Ux Ns -style
 date, but interpreted as local time rather
 than UTC.
 .It regex
 A regular expression match in extended
 .Tn POSIX
 regular expression syntax
 (much like egrep).
 The type specification can be optionally followed by
 .Ql /c
 for case-insensitive matches.
 The regular expression is always
 tested against the first
 .Ar N
 lines, where
 .Ar N
 is the given offset, thus it
 is only useful for (single-byte encoded) text.
 .Ql ^
 and
 .Ql $
 will match the beginning and end of individual lines, respectively,
 not beginning and end of file.
 .It search
 A literal string search starting at the given offset.
 It must be followed by
 .Li / Ns Aq Ar number
 which specifies how many matches shall be attempted (the range).
 This is suitable for searching larger binary expressions with variable
 offsets, using
 .Ql \e
 escapes for special characters.
 .El
 .El
 .Pp
 The numeric types may optionally be followed by
 .Em &
 and a numeric value,
 to specify that the value is to be AND'ed with the
 numeric value before any comparisons are done.
 Prepending a
 .Em u
 to the type indicates that ordered comparisons should be unsigned.
 .Bl -tag -width indent
 .It test
 The value to be compared with the value from the file.
 If the type is
 numeric, this value
 is specified in C form; if it is a string, it is specified as a C string
 with the usual escapes permitted (e.g.\& \en for new-line).
 .It ""
 Numeric values
 may be preceded by a character indicating the operation to be performed.
 It may be
 .Em = ,
 to specify that the value from the file must equal the specified value,
 .Em < ,
 to specify that the value from the file must be less than the specified
 value,
 .Em > ,
 to specify that the value from the file must be greater than the specified
 value,
 .Em & ,
 to specify that the value from the file must have set all of the bits
 that are set in the specified value,
 .Em ^ ,
 to specify that the value from the file must have clear any of the bits
 that are set in the specified value, or
 .Em ~ ,
 the value specified after is negated before tested, or
 .Em x ,
 to specify that any value will match.
 If the character is omitted,
 it is assumed to be
 .Em = .
 For all tests except
 .Dq string
 and
 .Dq regex ,
 operation
 .Em !\&
 specifies that the line matches if the test does
 .Em not
 succeed.
 .It ""
 Numeric values are specified in C form; e.g.\&
 .Em 13
 is decimal,
 .Em 013
 is octal, and
 .Em 0x13
 is hexadecimal.
 .It ""
 For string values, the byte string from the
 file must match the specified byte string.
 The operators
 .Em = ,
 .Em <
 and
 .Em >
 (but not
 .Em & )
 can be applied to strings.
 The length used for matching is that of the string argument
 in the magic file.
 This means that a line can match any string, and
 then presumably print that string, by doing
 .Em >\e0
 (because all strings are greater than the null string).
 .It message
 The message to be printed if the comparison succeeds.
 If the string
 contains a
 .Xr printf 3
 format specification, the value from the file (with any specified masking
 performed) is printed using the message as the format string.
 .El
 .Pp
 Some file formats contain additional information which is to be printed
 along with the file type or need additional tests to determine the true
 file type.
 These additional tests are introduced by one or more
 .Em >
 characters preceding the offset.
 The number of
 .Em >
 on the line indicates the level of the test; a line with no
 .Em >
 at the beginning is considered to be at level 0.
 Tests are arranged in a tree-like hierarchy:
 If a the test on a line at level
 .Em n
 succeeds, all following tests at level
 .Em n+1
 are performed, and the messages printed if the tests succeed, until a line
 with level
 .Em n
 (or less) appears.
 For more complex files, one can use empty messages to get just the
 "if/then" effect, in the following way:
 .Bd -literal -offset indent
 0      string   MZ
 >0x18  leshort  <0x40   MS-DOS executable
 >0x18  leshort  >0x3f   extended PC executable (e.g., MS Windows)
 .Ed
 .Pp
 Offsets do not need to be constant, but can also be read from the file
 being examined.
 If the first character following the last
 .Em >
 is a
 .Em \&(
 then the string after the parenthesis is interpreted as an indirect offset.
 That means that the number after the parenthesis is used as an offset in
 the file.
 The value at that offset is read, and is used again as an offset
 in the file.
 Indirect offsets are of the form:
 .Em (x[.[bslBSL]][+\-][y]) .
 The value of
 .Em x
 is used as an offset in the file.
 A byte, short or long is read at that offset
 depending on the
 .Em [bslBSLm]
 type specifier.
 The capitalized types interpret the number as a big endian value, whereas
 a small letter versions interpret the number as a little endian value;
 the
 .Em m
 type interprets the number as a middle endian (PDP-11) value.
 To that number the value of
 .Em y
 is added and the result is used as an offset in the file.
 The default type
 if one is not specified is long.
 .Pp
 That way variable length structures can be examined:
 .Bd -literal -offset indent
 # MS Windows executables are also valid MS-DOS executables
 0           string  MZ
 >0x18       leshort <0x40   MZ executable (MS-DOS)
 # skip the whole block below if it is not an extended executable
 >0x18       leshort >0x3f
 >>(0x3c.l)  string  PE\e0\e0  PE executable (MS-Windows)
 >>(0x3c.l)  string  LX\e0\e0  LX executable (OS/2)
 .Ed
 .Pp
 This strategy of examining has one drawback: You must make sure that
 you eventually print something, or users may get empty output (like, when
 there is neither PE\e0\e0 nor LE\e0\e0 in the above example).
 .Pp
 If this indirect offset cannot be used as-is, there are simple calculations
 possible: appending
 .Em [+-*/%&|^]<number>
 inside parentheses allows one to modify
 the value read from the file before it is used as an offset:
 .Bd -literal -offset indent
 # MS Windows executables are also valid MS-DOS executables
 0           string  MZ
 # sometimes, the value at 0x18 is less that 0x40 but there's still an
 # extended executable, simply appended to the file
 >0x18       leshort <0x40
 >>(4.s*512) leshort 0x014c  COFF executable (MS-DOS, DJGPP)
 >>(4.s*512) leshort !0x014c MZ executable (MS-DOS)
 .Ed
 .Pp
 Sometimes you do not know the exact offset as this depends on the length or
 position (when indirection was used before) of preceding fields.
 You can
 specify an offset relative to the end of the last uplevel field using
 .Em &
 as a prefix to the offset:
 .Bd -literal -offset indent
 0           string  MZ
 >0x18       leshort >0x3f
 >>(0x3c.l)  string  PE\e0\e0    PE executable (MS-Windows)
 # immediately following the PE signature is the CPU type
 >>>&0       leshort 0x14c     for Intel 80386
 >>>&0       leshort 0x184     for DEC Alpha
 .Ed
 .Pp
 Indirect and relative offsets can be combined:
 .Bd -literal -offset indent
 0             string  MZ
 >0x18         leshort <0x40
 >>(4.s*512)   leshort !0x014c MZ executable (MS-DOS)
 # if it's not COFF, go back 512 bytes and add the offset taken
 # from byte 2/3, which is yet another way of finding the start
 # of the extended executable
 >>>&(2.s-514) string  LE      LE executable (MS Windows VxD driver)
 .Ed
 .Pp
 Or the other way around:
 .Bd -literal -offset indent
 0                 string  MZ
 >0x18             leshort >0x3f
 >>(0x3c.l)        string  LE\e0\e0  LE executable (MS-Windows)
 # at offset 0x80 (-4, since relative offsets start at the end
 # of the uplevel match) inside the LE header, we find the absolute
 # offset to the code area, where we look for a specific signature
 >>>(&0x7c.l+0x26) string  UPX     \eb, UPX compressed
 .Ed
 .Pp
 Or even both!
 .Bd -literal -offset indent
 0                string  MZ
 >0x18            leshort >0x3f
 >>(0x3c.l)       string  LE\e0\e0 LE executable (MS-Windows)
 # at offset 0x58 inside the LE header, we find the relative offset
 # to a data area where we look for a specific signature
 >>>&(&0x54.l-3)  string  UNACE  \eb, ACE self-extracting archive
 .Ed
 .Pp
 Finally, if you have to deal with offset/length pairs in your file, even the
 second value in a parenthesed expression can be taken from the file itself,
 using another set of parentheses.
 Note that this additional indirect offset
 is always relative to the start of the main indirect offset.
 .Bd -literal -offset indent
 0                 string       MZ
 >0x18             leshort      >0x3f
 >>(0x3c.l)        string       PE\e0\e0 PE executable (MS-Windows)
 # search for the PE section called ".idata"...
 >>>&0xf4          search/0x140 .idata
 # ...and go to the end of it, calculated from start+length;
 # these are located 14 and 10 bytes after the section name
 >>>>(&0xe.l+(-4)) string       PK\e3\e4 \eb, ZIP self-extracting archive
 .Ed
 .Sh BUGS
 The formats
 .Em long ,
 .Em belong ,
 .Em lelong ,
 .Em melong ,
 .Em short ,
 .Em beshort ,
 .Em leshort ,
 .Em date ,
 .Em bedate ,
 .Em medate ,
 .Em ledate ,
 .Em beldate ,
 .Em leldate ,
 and
 .Em meldate
 are system-dependent; perhaps they should be specified as a number
 of bytes (2B, 4B, etc),
 since the files being recognized typically come from
 a system on which the lengths are invariant.
 .Pp
 If
 .Pa /usr/share/misc/magic
 is newer than
 .Pa /usr/share/misc/magic.mgc
 it is not used.
 Use the command:
 .Dq Li "cd /usr/share/misc && file -C -m magic"
 to rebuild.
 .Sh SEE ALSO
 .Xr file 1
 .\"
 .\" From: guy@sun.uucp (Guy Harris)
 .\" Newsgroups: net.bugs.usg
 .\" Subject: /etc/magic's format isn't well documented
 .\" Message-ID: <2752@sun.uucp>
 .\" Date: 3 Sep 85 08:19:07 GMT
 .\" Organization: Sun Microsystems, Inc.
 .\" Lines: 136
 .\"
 .\" Here's a manual page for the format accepted by the "file" made by adding
 .\" the changes I posted to the S5R2 version.
 .\"
 .\" Modified for Ian Darwin's version of the file command.
 .\" @(#)$Id: magic.man,v 1.30 2006/02/19 18:16:03 christos Exp $