[BACK]Return to file.1 CVS log [TXT][DIR] Up to [local] / src / usr.bin / file

Diff for /src/usr.bin/file/file.1 between version 1.35 and 1.36

version 1.35, 2015/02/15 22:26:45 version 1.36, 2015/04/24 16:24:11
Line 1 
Line 1 
 .\" $OpenBSD$  .\" $OpenBSD$
 .\" $FreeBSD: src/usr.bin/file/file.1,v 1.16 2000/03/01 12:19:39 sheldonh Exp $  .\" $FreeBSD: src/usr.bin/file/file.1,v 1.16 2000/03/01 12:19:39 sheldonh Exp $
 .\"  .\"
   .\" Copyright (c) 2015 Nicholas Marriott <nicm@openbsd.org>
 .\" Copyright (c) Ian F. Darwin 1986-1995.  .\" Copyright (c) Ian F. Darwin 1986-1995.
 .\" Software written by Ian F. Darwin and others;  .\" Software written by Ian F. Darwin and others;
 .\" maintained 1995-present by Christos Zoulas and others.  .\" maintained 1995-present by Christos Zoulas and others.
Line 36 
Line 37 
 .Sh SYNOPSIS  .Sh SYNOPSIS
 .Nm  .Nm
 .Bk -words  .Bk -words
 .Op Fl 0bCcehikLNnprsvz  .Op Fl bciLsW
 .Op Fl -help  .Ar
 .Op Fl -mime-encoding  
 .Op Fl -mime-type  
 .Op Fl F Ar separator  
 .Op Fl f Ar namefile  
 .Op Fl m Ar magicfiles  
 .Ar file  
 .Ek  .Ek
 .Sh DESCRIPTION  .Sh DESCRIPTION
 The  The
 .Nm  .Nm
 utility tests each argument in an attempt to classify it.  utility tests each argument and attempts to determine its type.
 There are three sets of tests, performed in this order:  Three sets of tests are performed:
 filesystem tests, magic tests, and language tests.  .Bl -enum -offset Ds
 The first test that succeeds causes the file type to be printed.  .It
   Filesystem tests, for example if a file is empty, or a special file such as a
   socket or named pipe (FIFO).
   .It
   .Dq Magic
   tests for data in particular fixed formats.
   These are loaded from the
   .Pa /etc/magic
   file (or
   .Pa ~/.magic
   instead if it exists).
   The file format is described in
   .Xr magic 5 .
   .It
   Tests for text files such as plain ASCII or C programming language files.
   .El
 .Pp  .Pp
 The type printed will usually contain one of the words  The first test which succeeds causes the file type to be printed.
   The type will often contain one of the words
 .Em text  .Em text
 (the file contains only  (contains only printing characters and is probably safe to read on an ASCII
 printing characters and a few common control  terminal),
 characters and is probably safe to read on an  
 ASCII terminal),  
 .Em executable  .Em executable
 (the file contains the result of compiling a program  (the file contains a compiled executable program)
 in a form understandable to some  
 .Ux  
 kernel or another),  
 or  or
 .Em data  .Em data
 meaning anything else (data is usually  meaning anything else.
 .Dq binary  
 or non-printable).  
 Exceptions are well-known file formats (core files, tar archives)  
 that are known to contain binary data.  
 When modifying magic files or the program itself, make sure to  
 .Em preserve these keywords .  
 Users depend on knowing that all the readable files in a directory  
 have the word  
 .Dq text  
 printed.  
 Don't do as Berkeley did and change  
 .Dq shell commands text  
 to  
 .Dq shell script .  
 .Pp  
 The filesystem 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's some sort of special file.  
 Any known file types,  
 such as sockets, symbolic links, and named pipes (FIFOs),  
 are intuited if they are defined in  
 the system header file  
 .In sys/stat.h .  
 .Pp  
 The magic tests are used to check for files with data in  
 particular fixed formats.  
 The canonical example of this is a binary executable (compiled program)  
 a.out file, whose format is defined in  
 .In elf.h ,  
 .In a.out.h ,  
 and possibly  
 .In exec.h  
 in the standard include directory.  
 These files have a  
 .Dq 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 a  
 .Dq magic  
 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 magic file  
 .Pa /etc/magic .  
 In addition, if  
 .Pa $HOME/.magic.mgc  
 or  
 .Pa $HOME/.magic  
 exists, it will be used in preference to the system magic files.  
 .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 text  
 because they will be mostly readable on nearly any terminal;  
 UTF-16 and EBCDIC are only  
 .Dq 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 Unix-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.\&  
 .In names.h )  
 that can appear anywhere in the first few blocks of a file.  
 For example, the keyword  
 .Em .br  
 indicates that the file is most likely a  
 troff input file, just as the keyword  
 .Em 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 data .  
 .Sh OPTIONS  .Sh OPTIONS
 .Bl -tag -width indent  .Bl -tag -width indent
 .It Fl 0 , -print0  .It Fl b
 Output a null character  Do not prepend filenames to output lines.
 .Sq \e0  .It Fl c
 after the end of the filename.  Print a summary of the parsed magic file, usually used for debugging.
 Nice to  .It Fl i , -mime , -mime-type
 .Xr cut 1  Causes the file command to output MIME type strings rather than the more
 the output.  traditional human-readable ones.
 This does not affect the separator which is still printed.  
 .It Fl b , -brief  
 Do not prepend filenames to output lines (brief mode).  
 .It Fl C , -compile  
 Write a  
 .Pa magic.mgc  
 output file that contains a pre-parsed version of the magic file or directory.  
 .It Fl c , -checking-printout  
 Cause a checking printout of the parsed form of the magic file.  
 This is usually used in conjunction with the  
 .Fl m  
 flag to debug a new magic file before installing it.  
 .It Fl e , -exclude Ar testname  
 Exclude the test named in  
 .Ar testname  
 from the list of tests made to determine the file type.  
 Valid test names are:  
 .Bl -tag -width compress  
 .It apptype  
 Check for  
 .Dv EMX  
 application type (only on EMX).  
 .It ascii  
 Check for various types of ASCII files.  
 .It compress  
 Don't look for, or inside, compressed files.  
 .It elf  
 Don't print elf details.  
 .It fortran  
 Don't look for fortran sequences inside ASCII files.  
 .It soft  
 Don't consult magic files.  
 .It tar  
 Don't examine tar files.  
 .It token  
 Don't look for known tokens inside ASCII files.  
 .It troff  
 Don't look for troff sequences inside ASCII files.  
 .El  
 .It Fl F , -separator Ar separator  
 Use the specified string as the separator between the filename and the  
 file result returned.  
 Defaults to  
 .Sq \&: .  
 .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  
 .Sq -  
 as a filename argument.  
 .It Fl h , -no-dereference  
 Causes symlinks not to be followed.  
 This is the default if the environment variable  
 .Dv POSIXLY_CORRECT  
 is not defined.  
 .It Fl -help  
 Print a help message and exit.  
 .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  Thus it may say
 .Dq text/plain charset=us-ascii  .Dq text/plain
 rather than  rather than
 .Dq ASCII text .  .Dq ASCII text .
 In order for this option to work,  .It Fl L
   Causes symlinks to be followed.
   .It Fl s
   Instructs
 .Nm  .Nm
 changes the way it handles files recognized by the command itself  to attempt to read all files, not only those which
 (such as many of the text file types, directories etc.),  
 and makes use of an alternative  
 .Dq magic  
 file.  
 See also  
 .Sx FILES ,  
 below.  
 .It Fl -mime-encoding , -mime-type  
 Like  
 .Fl i ,  
 but print only the specified element(s).  
 .It Fl k , -keep-going  
 Don't stop at the first match, keep going.  
 Subsequent matches will have the string  
 .Dq "\[rs]012\- "  
 prepended.  
 (If a newline is required, see the  
 .Fl r  
 option.)  
 .It Fl L , -dereference  
 Causes symlinks to be followed;  
 analogous to the option of the same name in  
 .Xr ls 1 .  
 This is the default if the environment variable  
 .Dv POSIXLY_CORRECT  
 is defined.  
 .It Fl m , -magic-file Ar magicfiles  
 Specify an alternate list of files and directories containing magic.  
 This can be a single item, or a colon-separated list.  
 If a compiled magic file is found alongside a file or directory,  
 it will be used instead.  
 .It Fl N , -no-pad  
 Don't pad filenames so that they align in the output.  
 .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 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  
 Don't 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  .Xr stat 2
 reports are ordinary files.  reports are ordinary files.
 This prevents problems, because reading special files may have peculiar  .It Fl W
 consequences.  Display warnings when parsing the magic file or applying its tests.
 Specifying the  Usually used for debugging.
 .Fl s  
 option causes  
 .Nm  
 to also read argument files which are block or character special files.  
 This is useful for determining the filesystem 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.  
 .El  .El
 .Sh ENVIRONMENT  
 The environment variable  
 .Dv MAGIC  
 can be used to set the default magic file name.  
 If that variable is set, then  
 .Nm  
 will not attempt to open  
 .Pa $HOME/.magic .  
 .Nm  
 adds  
 .Dq .mgc  
 to the value of this variable as appropriate.  
 The environment variable  
 .Dv POSIXLY_CORRECT  
 controls whether  
 .Nm  
 will attempt to follow symlinks or not.  
 If set, then  
 .Nm  
 follows symlinks; otherwise it does not.  
 This is also controlled by the  
 .Fl L  
 and  
 .Fl h  
 options.  
 .Sh FILES  .Sh FILES
 .Bl -tag -width /etc/magic -compact  .Bl -tag -width /etc/magic -compact
 .It Pa /etc/magic  .It Pa /etc/magic
 default list of magic numbers  default magic file
 .El  .El
 .Sh EXIT STATUS  .Sh EXIT STATUS
 .Ex -std file  .Ex -std file
 .Sh SEE ALSO  .Sh SEE ALSO
 .Xr hexdump 1 ,  
 .Xr od 1 ,  
 .Xr strings 1 ,  
 .Xr magic 5  .Xr magic 5
 .Sh STANDARDS CONFORMANCE  .Sh AUTHORS
 This program is believed to exceed the System V Interface Definition  
 of FILE(CMD), as near as one can determine from the vague language  
 contained therein.  
 Its behavior 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.  
 .\" URL: http://www.opengroup.org/onlinepubs/009695399/utilities/file.html  
 .Pp  
 The one significant difference  
 between this version and System V  
 is that this version treats any whitespace  
 as a delimiter, so that spaces in pattern strings must be escaped.  
 For example,  
 .Bd -literal -offset indent  
 \*(Gt10 string  language impress\       (imPRESS data)  
 .Ed  
 .Pp  
 in an existing magic file would have to be changed to  
 .Bd -literal -offset indent  
 \*(Gt10 string  language\e impress      (imPRESS data)  
 .Ed  
 .Pp  
 In addition, in this version, if a pattern string contains a backslash,  
 it must be escaped.  
 For example  
 .Bd -literal -offset indent  
 0       string          \ebegindata     Andrew Toolkit document  
 .Ed  
 .Pp  
 in an existing magic file would have to be changed to  
 .Bd -literal -offset indent  
 0       string          \e\ebegindata   Andrew Toolkit document  
 .Ed  
 .Pp  
 SunOS releases 3.2 and later from Sun Microsystems include a  
 .Nm  .Nm
 command derived from the System V one, but with some extensions.  commands have appeared in many previous versions of
 This version differs from Sun's only in minor ways.  .Ux .
 It includes the extension of the  This version was written by Nicholas Marriott for
 .Sq &  .Ox 5.8
 operator, used as,  to replace the previous version originally written by Ian Darwin.
 for example,  
 .Bd -literal -offset indent  
 \*(Gt16 long&0x7fffffff \*(Gt0          not stripped  
 .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 types.  
 This slowed the program down slightly but made it a lot more flexible.  
 .Pp  .Pp
 This program, based on the System V version,  There is a large number of contributors to the magic files; many are listed in
 was written by Ian Darwin  the source files.
 without looking at anybody else's source code.  
 .Pp  
 John Gilmore revised the code extensively, making it better than  
 the first version.  
 Geoff Collyer found several inadequacies  
 and provided some magic file entries.  
 Contributions by the `&' operator by Rob McMahon, 1989.  
 .Pp  
 Guy Harris, made many changes from 1993 to the present.  
 .Pp  
 Primary development and maintenance from 1990 to the present by  
 Christos Zoulas.  
 .Pp  
 Altered by Chris Lowth, 2000:  
 Handle the  
 .Fl i  
 option to output mime type strings, using an alternative  
 magic file and internal logic.  
 .Pp  
 Altered by Eric Fischer, July, 2000,  
 to identify character codes and attempt to identify the languages  
 of non-ASCII files.  
 .Pp  
 Altered by Reuben Thomas, 2007 to 2008, to improve MIME  
 support and merge MIME and non-MIME magic, support directories as well  
 as files of magic, apply many bug fixes and improve the build system.  
 .Pp  
 The list of contributors to the  
 .Dq magic  
 directory (magic files)  
 is too long to include here.  
 You know who you are; thank you.  
 Many contributors are listed in the source files.  
 .Sh BUGS  
 There must be a better way to automate the construction of the Magic  
 file from all the glop in Magdir.  
 What is it?  
 .Pp  
 .Nm  
 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  
 The list of keywords in  
 .Pa ascmagic  
 probably belongs in the Magic file.  
 This could be done by using some keyword like  
 .Sq *  
 for the offset value.  
 .Pp  
 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 From\  
 as first 5 chars of file) because  
 they are not as good as other guesses (e.g.\&  
 .Dq Newsgroups:  
 versus  
 .Dq Return-Path: ) .  
 Still, if the others don't pan out, it should be possible to use the  
 first guess.  
 .Pp  
 This manual page, and particularly this section, is too long.  
Legend:
Removed from v.1.35  
changed lines
  Added in v.1.36