Magic number file for the file command (UNIX)
The file command identifies the type of a file using, among other tests, a test for whether the file begins with a certain magic number. The file /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.
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:
The numeric types may optionally be followed by & and a numeric value, to specify that the value is to be ANDed with the numeric value before any comparisons are done. Prepending a u to the type indicates that ordered comparisons should be unsigned.
Numeric values may be preceded by a character indicating the operation to be performed:
If the character is omitted, it's assumed to be =.
Numeric values are specified in C form; e.g. 13 is decimal, 013 is octal, and 0x13 is hexadecimal.
For string values, the byte string from the file must match the specified byte string. The operators =, < and > (but not &) can be applied to strings. The length used for matching is that of the string argument in the /usr/share/misc/magic file. This means that a line can match any string, and then presumably print that string, by doing >\0 (because all strings are greater than the null string).
Some file formats contain additional information that's to be printed along with the file type. A line that begins with the character > indicates additional tests and messages to be printed. The number of > characters on the line indicates the level of the test; a line with no > at the beginning is considered to be at level 0. Each line at level n+1 is under the control of the line at level n most closely preceding it in the magic file.
If the test on a line at level n succeeds, the tests specified in all the subsequent lines at level n+1 are performed, and the messages printed if the tests succeed. The next line at level n terminates this.
If the first character following the last > is a (, 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: ((x[.[bsl]][+-][y]). The value of x is used as an offset in the file. A byte, short or long is read at that offset depending on the [bsl] type specifier. To that number the value of y is added and the result is used as an offset in the file. The default type if one isn't specified is long.
The formats long, belong, lelong, short, beshort, leshort, date, bedate, and ledate 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.
There's (currently) no support for specified-endian data to be used in indirect offsets.