1.\" $File: file.man,v 1.110 2014/11/28 02:46:39 christos Exp $ 2.Dd November 27, 2014 3.Dt FILE __CSECTION__ 4.Os 5.Sh NAME 6.Nm file 7.Nd determine file type 8.Sh SYNOPSIS 9.Nm 10.Bk -words 11.Op Fl bcEhiklLNnprsvz0 12.Op Fl Fl apple 13.Op Fl Fl mime-encoding 14.Op Fl Fl mime-type 15.Op Fl e Ar testname 16.Op Fl F Ar separator 17.Op Fl f Ar namefile 18.Op Fl m Ar magicfiles 19.Op Fl P Ar name=value 20.Ar 21.Ek 22.Nm 23.Fl C 24.Op Fl m Ar magicfiles 25.Nm 26.Op Fl Fl help 27.Sh DESCRIPTION 28This manual page documents version __VERSION__ of the 29.Nm 30command. 31.Pp 32.Nm 33tests each argument in an attempt to classify it. 34There are three sets of tests, performed in this order: 35filesystem tests, magic tests, and language tests. 36The 37.Em first 38test that succeeds causes the file type to be printed. 39.Pp 40The type printed will usually contain one of the words 41.Em text 42(the file contains only 43printing characters and a few common control 44characters and is probably safe to read on an 45.Dv ASCII 46terminal), 47.Em executable 48(the file contains the result of compiling a program 49in a form understandable to some 50.Tn UNIX 51kernel or another), 52or 53.Em data 54meaning anything else (data is usually 55.Dq binary 56or non-printable). 57Exceptions are well-known file formats (core files, tar archives) 58that are known to contain binary data. 59When modifying magic files or the program itself, make sure to 60.Em "preserve these keywords" . 61Users depend on knowing that all the readable files in a directory 62have the word 63.Dq text 64printed. 65Don't do as Berkeley did and change 66.Dq shell commands text 67to 68.Dq shell script . 69.Pp 70The filesystem tests are based on examining the return from a 71.Xr stat 2 72system call. 73The program checks to see if the file is empty, 74or if it's some sort of special file. 75Any known file types appropriate to the system you are running on 76(sockets, symbolic links, or named pipes (FIFOs) on those systems that 77implement them) 78are intuited if they are defined in the system header file 79.In sys/stat.h . 80.Pp 81The magic tests are used to check for files with data in 82particular fixed formats. 83The canonical example of this is a binary executable (compiled program) 84.Dv a.out 85file, whose format is defined in 86.In elf.h , 87.In a.out.h 88and possibly 89.In exec.h 90in the standard include directory. 91These files have a 92.Dq "magic number" 93stored in a particular place 94near the beginning of the file that tells the 95.Tn UNIX 96operating system 97that the file is a binary executable, and which of several types thereof. 98The concept of a 99.Dq "magic" 100has been applied by extension to data files. 101Any file with some invariant identifier at a small fixed 102offset into the file can usually be described in this way. 103The information identifying these files is read from the compiled 104magic file 105.Pa __MAGIC__.mgc , 106or the files in the directory 107.Pa __MAGIC__ 108if the compiled file does not exist. 109In addition, if 110.Pa $HOME/.magic.mgc 111or 112.Pa $HOME/.magic 113exists, it will be used in preference to the system magic files. 114.Pp 115If a file does not match any of the entries in the magic file, 116it is examined to see if it seems to be a text file. 117ASCII, ISO-8859-x, non-ISO 8-bit extended-ASCII character sets 118(such as those used on Macintosh and IBM PC systems), 119UTF-8-encoded Unicode, UTF-16-encoded Unicode, and EBCDIC 120character sets can be distinguished by the different 121ranges and sequences of bytes that constitute printable text 122in each set. 123If a file passes any of these tests, its character set is reported. 124ASCII, ISO-8859-x, UTF-8, and extended-ASCII files are identified 125as 126.Dq text 127because they will be mostly readable on nearly any terminal; 128UTF-16 and EBCDIC are only 129.Dq character data 130because, while 131they contain text, it is text that will require translation 132before it can be read. 133In addition, 134.Nm 135will attempt to determine other characteristics of text-type files. 136If the lines of a file are terminated by CR, CRLF, or NEL, instead 137of the Unix-standard LF, this will be reported. 138Files that contain embedded escape sequences or overstriking 139will also be identified. 140.Pp 141Once 142.Nm 143has determined the character set used in a text-type file, 144it will 145attempt to determine in what language the file is written. 146The language tests look for particular strings (cf. 147.In names.h ) 148that can appear anywhere in the first few blocks of a file. 149For example, the keyword 150.Em .br 151indicates that the file is most likely a 152.Xr troff 1 153input file, just as the keyword 154.Em struct 155indicates a C program. 156These tests are less reliable than the previous 157two groups, so they are performed last. 158The language test routines also test for some miscellany 159(such as 160.Xr tar 1 161archives). 162.Pp 163Any file that cannot be identified as having been written 164in any of the character sets listed above is simply said to be 165.Dq data . 166.Sh OPTIONS 167.Bl -tag -width indent 168.It Fl Fl apple 169Causes the file command to output the file type and creator code as 170used by older MacOS versions. The code consists of eight letters, 171the first describing the file type, the latter the creator. 172.It Fl b , Fl Fl brief 173Do not prepend filenames to output lines (brief mode). 174.It Fl C , Fl Fl compile 175Write a 176.Pa magic.mgc 177output file that contains a pre-parsed version of the magic file or directory. 178.It Fl c , Fl Fl checking-printout 179Cause a checking printout of the parsed form of the magic file. 180This is usually used in conjunction with the 181.Fl m 182flag to debug a new magic file before installing it. 183.It Fl E 184On filesystem errors (file not found etc), instead of handling the error 185as regular output as POSIX mandates and keep going, issue an error message 186and exit. 187.It Fl e , Fl Fl exclude Ar testname 188Exclude the test named in 189.Ar testname 190from the list of tests made to determine the file type. 191Valid test names are: 192.Bl -tag -width compress 193.It apptype 194.Dv EMX 195application type (only on EMX). 196.It ascii 197Various types of text files (this test will try to guess the text 198encoding, irrespective of the setting of the 199.Sq encoding 200option). 201.It encoding 202Different text encodings for soft magic tests. 203.It tokens 204Ignored for backwards compatibility. 205.It cdf 206Prints details of Compound Document Files. 207.It compress 208Checks for, and looks inside, compressed files. 209.It elf 210Prints ELF file details. 211.It soft 212Consults magic files. 213.It tar 214Examines tar files. 215.El 216.It Fl F , Fl Fl separator Ar separator 217Use the specified string as the separator between the filename and the 218file result returned. 219Defaults to 220.Sq \&: . 221.It Fl f , Fl Fl files-from Ar namefile 222Read the names of the files to be examined from 223.Ar namefile 224(one per line) 225before the argument list. 226Either 227.Ar namefile 228or at least one filename argument must be present; 229to test the standard input, use 230.Sq - 231as a filename argument. 232Please note that 233.Ar namefile 234is unwrapped and the enclosed filenames are processed when this option is 235encountered and before any further options processing is done. 236This allows one to process multiple lists of files with different command line 237arguments on the same 238.Nm 239invocation. 240Thus if you want to set the delimiter, you need to do it before you specify 241the list of files, like: 242.Dq Fl F Ar @ Fl f Ar namefile , 243instead of: 244.Dq Fl f Ar namefile Fl F Ar @ . 245.It Fl h , Fl Fl no-dereference 246option causes symlinks not to be followed 247(on systems that support symbolic links). 248This is the default if the environment variable 249.Dv POSIXLY_CORRECT 250is not defined. 251.It Fl i , Fl Fl mime 252Causes the file command to output mime type strings rather than the more 253traditional human readable ones. 254Thus it may say 255.Sq text/plain; charset=us-ascii 256rather than 257.Dq ASCII text . 258.It Fl Fl mime-type , Fl Fl mime-encoding 259Like 260.Fl i , 261but print only the specified element(s). 262.It Fl k , Fl Fl keep-going 263Don't stop at the first match, keep going. 264Subsequent matches will be 265have the string 266.Sq "\[rs]012\- " 267prepended. 268(If you want a newline, see the 269.Fl r 270option.) 271The magic pattern with the highest strength (see the 272.Fl l 273option) comes first. 274.It Fl l , Fl Fl list 275Shows a list of patterns and their strength sorted descending by 276.Xr magic 4 277strength 278which is used for the matching (see also the 279.Fl k 280option). 281.It Fl L , Fl Fl dereference 282option causes symlinks to be followed, as the like-named option in 283.Xr ls 1 284(on systems that support symbolic links). 285This is the default if the environment variable 286.Ev POSIXLY_CORRECT 287is defined. 288.It Fl m , Fl Fl magic-file Ar magicfiles 289Specify an alternate list of files and directories containing magic. 290This can be a single item, or a colon-separated list. 291If a compiled magic file is found alongside a file or directory, 292it will be used instead. 293.It Fl N , Fl Fl no-pad 294Don't pad filenames so that they align in the output. 295.It Fl n , Fl Fl no-buffer 296Force stdout to be flushed after checking each file. 297This is only useful if checking a list of files. 298It is intended to be used by programs that want filetype output from a pipe. 299.It Fl p , Fl Fl preserve-date 300On systems that support 301.Xr utime 3 302or 303.Xr utimes 2 , 304attempt to preserve the access time of files analyzed, to pretend that 305.Nm 306never read them. 307.It Fl P , Fl Fl parameter Ar name=value 308Set various parameter limits. 309.Bl -column "elf_phnum" "Default" "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX" -offset indent 310.It Sy "Name" Ta Sy "Default" Ta Sy "Explanation" 311.It Li indir Ta 15 Ta recursion limit for indirect magic 312.It Li name Ta 30 Ta use count limit for name/use magic 313.It Li elf_phnum Ta 128 Ta max ELF program sections processed 314.It Li elf_shnum Ta 32768 Ta max ELF sections processed 315.El 316.It Fl r , Fl Fl raw 317Don't translate unprintable characters to \eooo. 318Normally 319.Nm 320translates unprintable characters to their octal representation. 321.It Fl s , Fl Fl special-files 322Normally, 323.Nm 324only attempts to read and determine the type of argument files which 325.Xr stat 2 326reports are ordinary files. 327This prevents problems, because reading special files may have peculiar 328consequences. 329Specifying the 330.Fl s 331option causes 332.Nm 333to also read argument files which are block or character special files. 334This is useful for determining the filesystem types of the data in raw 335disk partitions, which are block special files. 336This option also causes 337.Nm 338to disregard the file size as reported by 339.Xr stat 2 340since on some systems it reports a zero size for raw disk partitions. 341.It Fl v , Fl Fl version 342Print the version of the program and exit. 343.It Fl z , Fl Fl uncompress 344Try to look inside compressed files. 345.It Fl 0 , Fl Fl print0 346Output a null character 347.Sq \e0 348after the end of the filename. 349Nice to 350.Xr cut 1 351the output. 352This does not affect the separator, which is still printed. 353.It Fl -help 354Print a help message and exit. 355.El 356.Sh FILES 357.Bl -tag -width __MAGIC__.mgc -compact 358.It Pa __MAGIC__.mgc 359Default compiled list of magic. 360.It Pa __MAGIC__ 361Directory containing default magic files. 362.El 363.Sh ENVIRONMENT 364The environment variable 365.Ev MAGIC 366can be used to set the default magic file name. 367If that variable is set, then 368.Nm 369will not attempt to open 370.Pa $HOME/.magic . 371.Nm 372adds 373.Dq Pa .mgc 374to the value of this variable as appropriate. 375However, 376.Pa file 377has to exist in order for 378.Pa file.mime 379to be considered. 380The environment variable 381.Ev POSIXLY_CORRECT 382controls (on systems that support symbolic links), whether 383.Nm 384will attempt to follow symlinks or not. 385If set, then 386.Nm 387follows symlink, otherwise it does not. 388This is also controlled by the 389.Fl L 390and 391.Fl h 392options. 393.Sh SEE ALSO 394.Xr magic __FSECTION__ , 395.Xr hexdump 1 , 396.Xr od 1 , 397.Xr strings 1 , 398.Xr fstyp 8 399.Sh STANDARDS CONFORMANCE 400This program is believed to exceed the System V Interface Definition 401of FILE(CMD), as near as one can determine from the vague language 402contained therein. 403Its behavior is mostly compatible with the System V program of the same name. 404This version knows more magic, however, so it will produce 405different (albeit more accurate) output in many cases. 406.\" URL: http://www.opengroup.org/onlinepubs/009695399/utilities/file.html 407.Pp 408The one significant difference 409between this version and System V 410is that this version treats any white space 411as a delimiter, so that spaces in pattern strings must be escaped. 412For example, 413.Bd -literal -offset indent 414\*[Gt]10 string language impress\ (imPRESS data) 415.Ed 416.Pp 417in an existing magic file would have to be changed to 418.Bd -literal -offset indent 419\*[Gt]10 string language\e impress (imPRESS data) 420.Ed 421.Pp 422In addition, in this version, if a pattern string contains a backslash, 423it must be escaped. 424For example 425.Bd -literal -offset indent 4260 string \ebegindata Andrew Toolkit document 427.Ed 428.Pp 429in an existing magic file would have to be changed to 430.Bd -literal -offset indent 4310 string \e\ebegindata Andrew Toolkit document 432.Ed 433.Pp 434SunOS releases 3.2 and later from Sun Microsystems include a 435.Nm 436command derived from the System V one, but with some extensions. 437This version differs from Sun's only in minor ways. 438It includes the extension of the 439.Sq \*[Am] 440operator, used as, 441for example, 442.Bd -literal -offset indent 443\*[Gt]16 long\*[Am]0x7fffffff \*[Gt]0 not stripped 444.Ed 445.Sh MAGIC DIRECTORY 446The magic file entries have been collected from various sources, 447mainly USENET, and contributed by various authors. 448Christos Zoulas (address below) will collect additional 449or corrected magic file entries. 450A consolidation of magic file entries 451will be distributed periodically. 452.Pp 453The order of entries in the magic file is significant. 454Depending on what system you are using, the order that 455they are put together may be incorrect. 456If your old 457.Nm 458command uses a magic file, 459keep the old magic file around for comparison purposes 460(rename it to 461.Pa __MAGIC__.orig ) . 462.Sh EXAMPLES 463.Bd -literal -offset indent 464$ file file.c file /dev/{wd0a,hda} 465file.c: C program text 466file: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), 467 dynamically linked (uses shared libs), stripped 468/dev/wd0a: block special (0/0) 469/dev/hda: block special (3/0) 470 471$ file -s /dev/wd0{b,d} 472/dev/wd0b: data 473/dev/wd0d: x86 boot sector 474 475$ file -s /dev/hda{,1,2,3,4,5,6,7,8,9,10} 476/dev/hda: x86 boot sector 477/dev/hda1: Linux/i386 ext2 filesystem 478/dev/hda2: x86 boot sector 479/dev/hda3: x86 boot sector, extended partition table 480/dev/hda4: Linux/i386 ext2 filesystem 481/dev/hda5: Linux/i386 swap file 482/dev/hda6: Linux/i386 swap file 483/dev/hda7: Linux/i386 swap file 484/dev/hda8: Linux/i386 swap file 485/dev/hda9: empty 486/dev/hda10: empty 487 488$ file -i file.c file /dev/{wd0a,hda} 489file.c: text/x-c 490file: application/x-executable 491/dev/hda: application/x-not-regular-file 492/dev/wd0a: application/x-not-regular-file 493 494.Ed 495.Sh HISTORY 496There has been a 497.Nm 498command in every 499.Dv UNIX since at least Research Version 4 500(man page dated November, 1973). 501The System V version introduced one significant major change: 502the external list of magic types. 503This slowed the program down slightly but made it a lot more flexible. 504.Pp 505This program, based on the System V version, 506was written by Ian Darwin 507.Aq ian@darwinsys.com 508without looking at anybody else's source code. 509.Pp 510John Gilmore revised the code extensively, making it better than 511the first version. 512Geoff Collyer found several inadequacies 513and provided some magic file entries. 514Contributions by the 515.Sq \*[Am] 516operator by Rob McMahon, 517.Aq cudcv@warwick.ac.uk , 5181989. 519.Pp 520Guy Harris, 521.Aq guy@netapp.com , 522made many changes from 1993 to the present. 5231989. 524.Pp 525Primary development and maintenance from 1990 to the present by 526Christos Zoulas 527.Aq christos@astron.com . 528.Pp 529Altered by Chris Lowth 530.Aq chris@lowth.com , 5312000: handle the 532.Fl i 533option to output mime type strings, using an alternative 534magic file and internal logic. 535.Pp 536Altered by Eric Fischer 537.Aq enf@pobox.com , 538July, 2000, 539to identify character codes and attempt to identify the languages 540of non-ASCII files. 541.Pp 542Altered by Reuben Thomas 543.Aq rrt@sc3d.org , 5442007-2011, to improve MIME support, merge MIME and non-MIME magic, 545support directories as well as files of magic, apply many bug fixes, 546update and fix a lot of magic, improve the build system, improve the 547documentation, and rewrite the Python bindings in pure Python. 548.Pp 549The list of contributors to the 550.Sq magic 551directory (magic files) 552is too long to include here. 553You know who you are; thank you. 554Many contributors are listed in the source files. 555.Sh LEGAL NOTICE 556Copyright (c) Ian F. Darwin, Toronto, Canada, 1986-1999. 557Covered by the standard Berkeley Software Distribution copyright; see the file 558COPYING in the source distribution. 559.Pp 560The files 561.Pa tar.h 562and 563.Pa is_tar.c 564were written by John Gilmore from his public-domain 565.Xr tar 1 566program, and are not covered by the above license. 567.Sh RETURN CODE 568.Nm 569returns 0 on success, and non-zero on error. 570.Sh BUGS 571.Pp 572Please report bugs and send patches to the bug tracker at 573.Pa http://bugs.gw.com/ 574or the mailing list at 575.Aq file@mx.gw.com 576(visit 577.Pa http://mx.gw.com/mailman/listinfo/file 578first to subscribe). 579.Sh TODO 580.Pp 581Fix output so that tests for MIME and APPLE flags are not needed all 582over the place, and actual output is only done in one place. 583This needs a design. 584Suggestion: push possible outputs on to a list, then pick the 585last-pushed (most specific, one hopes) value at the end, or 586use a default if the list is empty. 587This should not slow down evaluation. 588.Pp 589Continue to squash all magic bugs. 590See Debian BTS for a good source. 591.Pp 592Store arbitrarily long strings, for example for %s patterns, so that 593they can be printed out. 594Fixes Debian bug #271672. 595Would require more complex store/load code in apprentice. 596.Pp 597Add syntax for relative offsets after current level (Debian bug #466037). 598.Pp 599Make file -ki work, i.e. give multiple MIME types. 600.Pp 601Add a zip library so we can peek inside Office2007 documents to 602figure out what they are. 603.Pp 604Add an option to print URLs for the sources of the file descriptions. 605.Pp 606Combine script searches and add a way to map executable names to MIME 607types (e.g. have a magic value for !:mime which causes the resulting 608string to be looked up in a table). 609This would avoid adding the same magic repeatedly for each new 610hash-bang interpreter. 611.Pp 612Fix 613.Dq name 614and 615.Dq use 616to check for consistency at compile time (duplicate 617.Dq name , 618.Dq use 619pointing to undefined 620.Dq name 621). 622Make 623.Dq name 624/ 625.Dq use 626more efficient by keeping a sorted list of names. 627Special-case ^ to flip endianness in the parser so that it does not 628have to be escaped, and document it. 629.Sh AVAILABILITY 630You can obtain the original author's latest version by anonymous FTP 631on 632.Pa ftp.astron.com 633in the directory 634.Pa /pub/file/file-X.YZ.tar.gz . 635