File

File is the open source and free implementation of the 'file' command used on almost every UNIX-like operating system, such as Linux and BSD. The 'file' command, if you're not familiar with it, is a command-line program that tells you what kind of data a file contains.

The original 'file' command shipped with Bell Labs UNIX, but was unavailable in source form to the masses before this re-implementation. Like any other command-line program, File must be used through any terminal emulator application.

Command-line options

File comes with various command-line options that will help you to classify files on your GNU/Linux system better and faster. They can be viewed at a glance by running the ‘file --help’ command in a terminal emulator app.

Among the most useful command-line options, we can mention support for using a specific file as a color-separated list of magic number files, support for processing compressed files, as well as support for outputting MIME type strings, MIME encoding or Apple type/creator.

Additionally, the user will be able to list magic strength, to enable following of symlinks, to terminate filenames with ASCII NUL, to preserve access time on files, to treat special files as ordinary ones, and to compile a file that is specified by the -m (--magic-file LIST) option.

Under the hood and supported OSes

File is a cross-platform software written entirely in the C programming language, which means that it is fast and supports all known GNU/Linux operating systems, as well as various BSD distributions, including FreeBSD, OpenBSD and NetBSD. Darwin/Mac OS X is also supported.

The program is distributed as a universal source tarball (TAR archive) that can be easily installed on any 32-bit or 64-bit system. To install it, simply download the archive, save it on your PC, unpack it, open a terminal app and navigate to the location of the extracted archive files (e.g. cd /home/softoware/file-5.20), run the ‘./configure && make’ command, and then execute the ‘make install’ or ‘sudo make install’ command.

What is new in this release:

• This is a bug fix release; most of the fixes were driven by oss-fuzz findings, and the reason for the release now is that there have been no new reports for more than two weeks. Most of the problems found by oss-fuzz were in...
• cdf parsing (out of bounds reads)
• string/regex matching (uninitialized memory reads)
• Other misc fixes:
• remove trailing spaces from magic files
• refactor is_tar

What is new in version 5.33:

• This is a bug fix release; most of the fixes were driven by oss-fuzz findings, and the reason for the release now is that there have been no new reports for more than two weeks. Most of the problems found by oss-fuzz were in...
• cdf parsing (out of bounds reads)
• string/regex matching (uninitialized memory reads)
• Other misc fixes:
• remove trailing spaces from magic files
• refactor is_tar

What is new in version :

• This is a bug fix release; most of the fixes were driven by oss-fuzz findings, and the reason for the release now is that there have been no new reports for more than two weeks. Most of the problems found by oss-fuzz were in...
• cdf parsing (out of bounds reads)
• string/regex matching (uninitialized memory reads)
• Other misc fixes:
• remove trailing spaces from magic files
• refactor is_tar

What is new in version 5.31:

• This is a bug fix release; most of the fixes were driven by oss-fuzz findings, and the reason for the release now is that there have been no new reports for more than two weeks. Most of the problems found by oss-fuzz were in...
• cdf parsing (out of bounds reads)
• string/regex matching (uninitialized memory reads)
• Other misc fixes:
• remove trailing spaces from magic files
• refactor is_tar

What is new in version 5.30:

• der getlength overflow (Jonas Wagner)
• multiple magic file load failure (Christoph Biedl)
• CDF parsing improvements (Guy Helmer)
• Add support for signed indirect offsets
• cat /dev/null | file - should print empty (Christoph Biedl)
• Bump string size from 64 to 96.
• PR/556: Fix separators on annotations.

What is new in version 5.29:

• der getlength overflow (Jonas Wagner)
• multiple magic file load failure (Christoph Biedl)
• CDF parsing improvements (Guy Helmer)
• Add support for signed indirect offsets
• cat /dev/null | file - should print empty (Christoph Biedl)
• Bump string size from 64 to 96.
• PR/556: Fix separators on annotations.

What is new in version 5.28:

• Features:
• the number of bytes read from files to determine the file type is now configurable (-P bytes=)
• add experimental DER parser ("der" softmagic field type)
• add NUL terminated output format -00
• decompression code now reports decompression errors
• add zlib support if the OS has it
• Bugfixes:
• fix ID3 indirect parsing
• provide dprintf(3) for the OS's that don't have it
• delete REG_STARTEND code is not working as expected
• forking for decompression was broken with magic_buffer

What is new in version 5.25:

• add a limit to the length of regex searches
• fix problems with --parameter
• Windows fixes PR/466

What is new in version 5.24:

• fix parsing of the mime long options
• adjust the FIT map magic to work after the date bug fixes in 5.23

What is new in version 5.22:

• Bugs:
• restructure elf note printing to avoid repeated messages
• add note limit, suggested by Alexander Cherepanov
• Bail out on partial pread()'s (Alexander Cherepanov)
• Fix incorrect bounds check in file_printable (Alexander Cherepanov)
• PR/405: ignore SIGPIPE from uncompress programs
• change printable -> file_printable and use it in more places for safety
• Features:
• add indirect relative for TIFF/Exif
• in ELF, instead of "(uses dynamic libraries)" when PT_INTERP is present print the interpreter name.

What is new in version 5.20:

• 1.sm.webp: RIFF (little-endian) data, Web/P image, VP8 encoding, 320x214, Scaling: [none]x[none], YUV color, decoders should clamp
• 1.webp: RIFF (little-endian) data, Web/P image, VP8 encoding, 550x368, Scaling: [none]x[none], YUV color, decoders should clamp
• CCITT_1.TIF: TIFF image data, little-endian, direntries=17, height=2376, bps=1, compression=bi-level group 4, PhotometricIntepretation=WhiteIsZero, name=Standard Input, description=converted PBM file, width=1728
• FLOWER.jpg: JPEG image data, baseline, precision 8, 2592x3904, frames 3
• FLOWER.wdp: JPEG-XR Image, hard tiling, spatial xform=TL, short header, 2592x3904, bitdepth=5-6-5, colorfmt=YONLY
• MARBLES.TIF: TIFF image data, little-endian, direntries=16, height=1001, bps=206, compression=LZW, PhotometricIntepretation=RGB, width=1419
• There are also a few bugs fixed from the previous releases, so this would be the recommended release to use...

What is new in version 5.17:

• Count recursion levels through indirect magic
• Prevent infinite recursion on files with indirect offsets of 0
• Add -E flag that makes file print filesystem errors to stderr and exit.
• mime printing could print results from multiple magic entries if there were multiple matches.
• In some cases overflow was not detected when computing offsets in softmagic.
• Use strcasestr() to for cdf strings
• Reset to the "C" locale while doing regex operations, or case insensitive comparisons; this is provisional

What is new in version 5.16:

• Always leave magic file loaded, don't unload for magic_check, etc.
• Fix default encoding to binary instead of unknown which broke recently
• Handle empty and one byte files, less specially so that --mime-encoding does not break completely
• Fix erroneous non-zero exit code from non-existant file and message
• Add CDF MSI file detection (Guy Helmer)