{
    "mode": "man",
    "parameter": "magic",
    "section": "5",
    "url": "https://www.chedong.com/phpMan.php/man/magic/5/json",
    "generated": "2026-06-02T18:07:10Z",
    "sections": {
        "NAME": {
            "content": "magic — file command's magic pattern file\n",
            "subsections": []
        },
        "DESCRIPTION": {
            "content": "This manual page documents the format of magic files as used by the file(1) command, version\n5.41.  The file(1) command identifies the type of a file using, among other tests, a test for\nwhether the file contains certain “magic patterns”.  The database of these “magic patterns” is\nusually located in a binary file in /usr/share/misc/magic.mgc or a directory of source text\nmagic pattern fragment files in /usr/share/misc/magic.  The database specifies what patterns\nare to be tested for, what message or MIME type to print if a particular pattern is found, and\nadditional information to extract from the file.\n\nThe format of the source fragment files that are used to build this database is as follows:\nEach line of a fragment file specifies a test to be performed.  A test compares the data start‐\ning at a particular offset in the file with a byte value, a string or a numeric value.  If the\ntest succeeds, a message is printed.  The line consists of the following fields:\n\noffset   A number specifying the offset (in bytes) into the file of the data which is to be\ntested.  This offset can be a negative number if it is:\n••   The first direct offset of the magic entry (at continuation level 0), in which\ncase it is interpreted an offset from end end of the file going backwards.  This\nworks only when a file descriptor to the file is available and it is a regular\nfile.\n••   A continuation offset relative to the end of the last up-level field (&).\n\ntype     The type of the data to be tested.  The possible values are:\n\nbyte        A one-byte value.\n\nshort       A two-byte value in this machine's native byte order.\n\nlong        A four-byte value in this machine's native byte order.\n\nquad        An eight-byte value in this machine's native byte order.\n\nfloat       A 32-bit single precision IEEE floating point number in this machine's na‐\ntive byte order.\n\ndouble      A 64-bit double precision IEEE floating point number in this machine's na‐\ntive byte order.\n\nstring      A string of bytes.  The string type specification can be optionally fol‐\nlowed by /[WwcCtbTf]*.  The “W” flag compacts whitespace in the target,\nwhich must contain at least one whitespace character.  If the magic has n\nconsecutive blanks, the target needs at least n consecutive blanks to\nmatch.  The “w” flag treats every blank in the magic as an optional blank.\nThe “f” flags requires that the matched string is a full word, not a par‐\ntial word match.  The “c” flag specifies case insensitive matching: lower\ncase characters in the magic match both lower and upper case characters in\nthe target, whereas upper case characters in the magic only match upper\ncase characters in the target.  The “C” flag specifies case insensitive\nmatching: upper case characters in the magic match both lower and upper\ncase characters in the target, whereas lower case characters in the magic\nonly match upper case characters in the target.  To do a complete case in‐\nsensitive match, specify both “c” and “C”.  The “t” flag forces the test\nto be done for text files, while the “b” flag forces the test to be done\nfor binary files.  The “T” flag causes the string to be trimmed, i.e.\nleading and trailing whitespace is deleted before the string is printed.\n\npstring     A Pascal-style string where the first byte/short/int is interpreted as the\nunsigned length.  The length defaults to byte and can be specified as a\nmodifier.  The following modifiers are supported:\nB  A byte length (default).\nH  A 2 byte big endian length.\nh  A 2 byte little endian length.\nL  A 4 byte big endian length.\nl  A 4 byte little endian length.\nJ  The length includes itself in its count.\nThe string is not NUL terminated.  “J” is used rather than the more valu‐\nable “I” because this type of length is a feature of the JPEG format.\n\ndate        A four-byte value interpreted as a UNIX date.\n\nqdate       An eight-byte value interpreted as a UNIX date.\n\nldate       A four-byte value interpreted as a UNIX-style date, but interpreted as lo‐\ncal time rather than UTC.\n\nqldate      An eight-byte value interpreted as a UNIX-style date, but interpreted as\nlocal time rather than UTC.\n\nqwdate      An eight-byte value interpreted as a Windows-style date.\n\nbeid3       A 32-bit ID3 length in big-endian byte order.\n\nbeshort     A two-byte value in big-endian byte order.\n\nbelong      A four-byte value in big-endian byte order.\n\nbequad      An eight-byte value in big-endian byte order.\n\nbefloat     A 32-bit single precision IEEE floating point number in big-endian byte\norder.\n\nbedouble    A 64-bit double precision IEEE floating point number in big-endian byte\norder.\n\nbedate      A four-byte value in big-endian byte order, interpreted as a Unix date.\n\nbeqdate     An eight-byte value in big-endian byte order, interpreted as a Unix date.\n\nbeldate     A four-byte value in big-endian byte order, interpreted as a UNIX-style\ndate, but interpreted as local time rather than UTC.\n\nbeqldate    An eight-byte value in big-endian byte order, interpreted as a UNIX-style\ndate, but interpreted as local time rather than UTC.\n\nbeqwdate    An eight-byte value in big-endian byte order, interpreted as a Windows-\nstyle date.\n\nbestring16  A two-byte unicode (UCS16) string in big-endian byte order.\n\nleid3       A 32-bit ID3 length in little-endian byte order.\n\nleshort     A two-byte value in little-endian byte order.\n\nlelong      A four-byte value in little-endian byte order.\n\nlequad      An eight-byte value in little-endian byte order.\n\nlefloat     A 32-bit single precision IEEE floating point number in little-endian byte\norder.\n\nledouble    A 64-bit double precision IEEE floating point number in little-endian byte\norder.\n\nledate      A four-byte value in little-endian byte order, interpreted as a UNIX date.\n\nleqdate     An eight-byte value in little-endian byte order, interpreted as a UNIX\ndate.\n\nleldate     A four-byte value in little-endian byte order, interpreted as a UNIX-style\ndate, but interpreted as local time rather than UTC.\n\nleqldate    An eight-byte value in little-endian byte order, interpreted as a UNIX-\nstyle date, but interpreted as local time rather than UTC.\n\nleqwdate    An eight-byte value in little-endian byte order, interpreted as a Windows-\nstyle date.\n\nlestring16  A two-byte unicode (UCS16) string in little-endian byte order.\n\nmelong      A four-byte value in middle-endian (PDP-11) byte order.\n\nmedate      A four-byte value in middle-endian (PDP-11) byte order, interpreted as a\nUNIX date.\n\nmeldate     A four-byte value in middle-endian (PDP-11) byte order, interpreted as a\nUNIX-style date, but interpreted as local time rather than UTC.\n\nindirect    Starting at the given offset, consult the magic database again.  The off‐\nset of the indirect magic is by default absolute in the file, but one can\nspecify /r to indicate that the offset is relative from the beginning of\nthe entry.\n\nname        Define a “named” magic instance that can be called from another use magic\nentry, like a subroutine call.  Named instance direct magic offsets are\nrelative to the offset of the previous matched entry, but indirect offsets\nare relative to the beginning of the file as usual.  Named magic entries\nalways match.\n\nuse         Recursively call the named magic starting from the current offset.  If the\nname of the referenced begins with a ^ then the endianness of the magic is\nswitched; if the magic mentioned leshort for example, it is treated as\nbeshort and vice versa.  This is useful to avoid duplicating the rules for\ndifferent endianness.\n\nregex       A regular expression match in extended POSIX regular expression syntax\n(like egrep).  Regular expressions can take exponential time to process,\nand their performance is hard to predict, so their use is discouraged.\nWhen used in production environments, their performance should be care‐\nfully checked.  The size of the string to search should also be limited by\nspecifying /<length>, to avoid performance issues scanning long files.\nThe type specification can also be optionally followed by /[c][s][l].  The\n“c” flag makes the match case insensitive, while the “s” flag update the\noffset to the start offset of the match, rather than the end.  The “l”\nmodifier, changes the limit of length to mean number of lines instead of a\nbyte count.  Lines are delimited by the platforms native line delimiter.\nWhen a line count is specified, an implicit byte count also computed as‐\nsuming each line is 80 characters long.  If neither a byte or line count\nis specified, the search is limited automatically to 8KiB.  ^ and $ match\nthe beginning and end of individual lines, respectively, not beginning and\nend of file.\n\nsearch      A literal string search starting at the given offset.  The same modifier\nflags can be used as for string patterns.  The search expression must con‐\ntain the range in the form /number, that is the number of positions at\nwhich the match will be attempted, starting from the start offset.  This\nis suitable for searching larger binary expressions with variable offsets,\nusing \\ escapes for special characters.  The order of modifier and number\nis not relevant.\n\ndefault     This is intended to be used with the test x (which is always true) and it\nhas no type.  It matches when no other test at that continuation level has\nmatched before.  Clearing that matched tests for a continuation level, can\nbe done using the clear test.\n\nclear       This test is always true and clears the match flag for that continuation\nlevel.  It is intended to be used with the default test.\n\nder         Parse the file as a DER Certificate file.  The test field is used as a der\ntype that needs to be matched.  The DER types are: eoc, bool, int,\nbitstr, octetstr, null, objid, objdesc, ext, real, enum, embed,\nutf8str, reloid, time, res2, seq, set, numstr, prtstr, t61str,\nvidstr, ia5str, utctime, gentime, grstr, visstr, genstr, univstr,\ncharstr, bmpstr, date, tod, datetime, duration, oid-iri, rel-oid-iri.\nThese types can be followed by an optional numeric size, which indicates\nthe field width in bytes.\n\nguid        A Globally Unique Identifier, parsed and printed as XXXXXXXX-XXXX-XXXX-\nXXXX-XXXXXXXXXXXX.  It's format is a string.\n\noffset      This is a quad value indicating the current offset of the file.  It can be\nused to determine the size of the file or the magic buffer.  For example\nthe magic entries:\n\n-0      offset  x       this file is %lld bytes\n-0      offset  <=100   must be more than 100 \\\nbytes and is only %lld\n\nFor compatibility with the Single UNIX Standard, the type specifiers dC and d1 are\nequivalent to byte, the type specifiers uC and u1 are equivalent to ubyte, the type\nspecifiers dS and d2 are equivalent to short, the type specifiers uS and u2 are equiv‐\nalent to ushort, the type specifiers dI, dL, and d4 are equivalent to long, the type\nspecifiers uI, uL, and u4 are equivalent to ulong, the type specifier d8 is equivalent\nto quad, the type specifier u8 is equivalent to uquad, and the type specifier s is\nequivalent to string.  In addition, the type specifier dQ is equivalent to quad and\nthe type specifier uQ is equivalent to uquad.\n\nEach top-level magic pattern (see below for an explanation of levels) is classified as\ntext or binary according to the types used.  Types “regex” and “search” are classified\nas text tests, unless non-printable characters are used in the pattern.  All other\ntests are classified as binary.  A top-level pattern is considered to be a test text\nwhen all its patterns are text patterns; otherwise, it is considered to be a binary\npattern.  When matching a file, binary patterns are tried first; if no match is found,\nand the file looks like text, then its encoding is determined and the text patterns\nare tried.\n\nThe numeric types may optionally be followed by & and a numeric value, to specify that\nthe value is to be AND'ed with the numeric value before any comparisons are done.\nPrepending a u to the type indicates that ordered comparisons should be unsigned.\n\ntest     The value to be compared with the value from the file.  If the type is numeric, this\nvalue is specified in C form; if it is a string, it is specified as a C string with\nthe usual escapes permitted (e.g. \\n for new-line).\n\nNumeric values may be preceded by a character indicating the operation to be per‐\nformed.  It may be =, to specify that the value from the file must equal the specified\nvalue, <, to specify that the value from the file must be less than the specified\nvalue, >, to specify that the value from the file must be greater than the specified\nvalue, &, to specify that the value from the file must have set all of the bits that\nare set in the specified value, ^, to specify that the value from the file must have\nclear any of the bits that are set in the specified value, or ~, the value specified\nafter is negated before tested.  x, to specify that any value will match.  If the\ncharacter is omitted, it is assumed to be =.  Operators &, ^, and ~ don't work with\nfloats and doubles.  The operator ! specifies that the line matches if the test does\nnot succeed.\n\nNumeric values are specified in C form; e.g.  13 is decimal, 013 is octal, and 0x13 is\nhexadecimal.\n\nNumeric operations are not performed on date types, instead the numeric value is in‐\nterpreted as an offset.\n\nFor string values, the string from the file must match the specified string.  The op‐\nerators =, < and > (but not &) can be applied to strings.  The length used for match‐\ning is that of the string argument in the magic file.  This means that a line can\nmatch any non-empty string (usually used to then print the string), with >\\0 (because\nall non-empty strings are greater than the empty string).\n\nDates are treated as numerical values in the respective internal representation.\n\nThe special test x always evaluates to true.\n\nmessage  The message to be printed if the comparison succeeds.  If the string contains a\nprintf(3) format specification, the value from the file (with any specified masking\nperformed) is printed using the message as the format string.  If the string begins\nwith “\\b”, the message printed is the remainder of the string with no whitespace added\nbefore it: multiple matches are normally separated by a single space.\n\nAn APPLE 4+4 character APPLE creator and type can be specified as:\n\n!:apple CREATYPE\n\nA MIME type is given on a separate line, which must be the next non-blank or comment line after\nthe magic line that identifies the file type, and has the following format:\n\n!:mime  MIMETYPE\n\ni.e. the literal string “!:mime” followed by the MIME type.\n\nAn optional strength can be supplied on a separate line which refers to the current magic de‐\nscription using the following format:\n\n!:strength OP VALUE\n\nThe operand OP can be: +, -, *, or / and VALUE is a constant between 0 and 255.  This constant\nis applied using the specified operand to the currently computed default magic strength.\n\nSome file formats contain additional information which is to be printed along with the file\ntype or need additional tests to determine the true file type.  These additional tests are in‐\ntroduced by one or more > characters preceding the offset.  The number of > on the line indi‐\ncates the level of the test; a line with no > at the beginning is considered to be at level 0.\nTests are arranged in a tree-like hierarchy: if the test on a line at level n succeeds, all\nfollowing tests at level n+1 are performed, and the messages printed if the tests succeed, un‐\ntil a line with level n (or less) appears.  For more complex files, one can use empty messages\nto get just the \"if/then\" effect, in the following way:\n\n0      string   MZ\n>0x18  leshort  <0x40   MS-DOS executable\n>0x18  leshort  >0x3f   extended PC executable (e.g., MS Windows)\n\nOffsets do not need to be constant, but can also be read from the file being examined.  If the\nfirst character following the last > is a ( then the string after the parenthesis is inter‐\npreted as an indirect offset.  That means that the number after the parenthesis is used as an\noffset in the file.  The value at that offset is read, and is used again as an offset in the\nfile.  Indirect offsets are of the form: (( x [[.,][bBcCeEfFgGhHiIlmsSqQ]][+-][ y ]).  The\nvalue of x is used as an offset in the file.  A byte, id3 length, short or long is read at that\noffset depending on the [bBcCeEfFgGhHiIlmsSqQ] type specifier.  The value is treated as signed\nif “”, is specified or unsigned if “”.  is specified.  The capitalized types interpret the num‐\nber as a big endian value, whereas the small letter versions interpret the number as a little\nendian value; the m type interprets the number as a middle endian (PDP-11) value.  To that num‐\nber the value of y is added and the result is used as an offset in the file.  The default type\nif one is not specified is long.  The following types are recognized:\n\nType    Sy Mnemonic   Sy Endian Sy Size\nbcBc    Byte/Char     N/A       1\nefg     Double        Little    8\nEFG     Double        Big       8\nhs      Half/Short    Little    2\nHS      Half/Short    Big       2\ni       ID3           Little    4\nI       ID3           Big       4\nm       Middle        Middle    4\nq       Quad          Little    8\nQ       Quad          Big       8\n\nThat way variable length structures can be examined:\n\n# MS Windows executables are also valid MS-DOS executables\n0           string  MZ\n>0x18       leshort <0x40   MZ executable (MS-DOS)\n# skip the whole block below if it is not an extended executable\n>0x18       leshort >0x3f\n>>(0x3c.l)  string  PE\\0\\0  PE executable (MS-Windows)\n>>(0x3c.l)  string  LX\\0\\0  LX executable (OS/2)\n\nThis strategy of examining has a drawback: you must make sure that you eventually print some‐\nthing, or users may get empty output (such as when there is neither PE\\0\\0 nor LE\\0\\0 in the\nabove example).\n\nIf this indirect offset cannot be used directly, simple calculations are possible: appending\n[+-*/%&|^]number inside parentheses allows one to modify the value read from the file before it\nis used as an offset:\n\n# MS Windows executables are also valid MS-DOS executables\n0           string  MZ\n# sometimes, the value at 0x18 is less that 0x40 but there's still an\n# extended executable, simply appended to the file\n>0x18       leshort <0x40\n>>(4.s*512) leshort 0x014c  COFF executable (MS-DOS, DJGPP)\n>>(4.s*512) leshort !0x014c MZ executable (MS-DOS)\n\nSometimes you do not know the exact offset as this depends on the length or position (when in‐\ndirection was used before) of preceding fields.  You can specify an offset relative to the end\nof the last up-level field using ‘&’ as a prefix to the offset:\n\n0           string  MZ\n>0x18       leshort >0x3f\n>>(0x3c.l)  string  PE\\0\\0    PE executable (MS-Windows)\n# immediately following the PE signature is the CPU type\n>>>&0       leshort 0x14c     for Intel 80386\n>>>&0       leshort 0x184     for DEC Alpha\n\nIndirect and relative offsets can be combined:\n\n0             string  MZ\n>0x18         leshort <0x40\n>>(4.s*512)   leshort !0x014c MZ executable (MS-DOS)\n# if it's not COFF, go back 512 bytes and add the offset taken\n# from byte 2/3, which is yet another way of finding the start\n# of the extended executable\n>>>&(2.s-514) string  LE      LE executable (MS Windows VxD driver)\n\nOr the other way around:\n\n0                 string  MZ\n>0x18             leshort >0x3f\n>>(0x3c.l)        string  LE\\0\\0  LE executable (MS-Windows)\n# at offset 0x80 (-4, since relative offsets start at the end\n# of the up-level match) inside the LE header, we find the absolute\n# offset to the code area, where we look for a specific signature\n>>>(&0x7c.l+0x26) string  UPX     \\b, UPX compressed\n\nOr even both!\n\n0                string  MZ\n>0x18            leshort >0x3f\n>>(0x3c.l)       string  LE\\0\\0 LE executable (MS-Windows)\n# at offset 0x58 inside the LE header, we find the relative offset\n# to a data area where we look for a specific signature\n>>>&(&0x54.l-3)  string  UNACE  \\b, ACE self-extracting archive\n\nIf you have to deal with offset/length pairs in your file, even the second value in a parenthe‐\nsized expression can be taken from the file itself, using another set of parentheses.  Note\nthat this additional indirect offset is always relative to the start of the main indirect off‐\nset.\n\n0                 string       MZ\n>0x18             leshort      >0x3f\n>>(0x3c.l)        string       PE\\0\\0 PE executable (MS-Windows)\n# search for the PE section called \".idata\"...\n>>>&0xf4          search/0x140 .idata\n# ...and go to the end of it, calculated from start+length;\n# these are located 14 and 10 bytes after the section name\n>>>>(&0xe.l+(-4)) string       PK\\3\\4 \\b, ZIP self-extracting archive\n\nIf you have a list of known values at a particular continuation level, and you want to provide\na switch-like default case:\n\n# clear that continuation level match\n>18     clear\n>18     lelong  1       one\n>18     lelong  2       two\n>18     default x\n# print default match\n>>18    lelong  x       unmatched 0x%x\n",
            "subsections": []
        },
        "SEE ALSO": {
            "content": "file(1) - the command that reads this file.\n",
            "subsections": []
        },
        "BUGS": {
            "content": "The formats long, belong, lelong, melong, short, beshort, and leshort do not depend on the\nlength of the C data types short and long on the platform, even though the Single UNIX Specifi‐\ncation implies that they do.  However, as OS X Mountain Lion has passed the Single UNIX Speci‐\nfication validation suite, and supplies a version of file(1) in which they do not depend on the\nsizes of the C data types and that is built for a 64-bit environment in which long is 8 bytes\nrather than 4 bytes, presumably the validation suite does not test whether, for example long\nrefers to an item with the same size as the C data type long.  There should probably be type\nnames int8, uint8, int16, uint16, int32, uint32, int64, and uint64, and specified-byte-order\nvariants of them, to make it clearer that those types have specified widths.\n\nBSD                               May 9, 2021                              BSD",
            "subsections": []
        }
    },
    "summary": "magic — file command's magic pattern file",
    "flags": [],
    "examples": [],
    "see_also": [
        {
            "name": "file",
            "section": "1",
            "url": "https://www.chedong.com/phpMan.php/man/file/1/json"
        }
    ]
}