Detecting the MIME Type for a File

Applications can detect the MIME type of a file as follows:

  1. The application uses file content sniffers to search for a particular pattern in the file. A file content sniffer associates a specific pattern in a file with a MIME type. If the application finds a match for the pattern, the MIME type associated with the pattern is the MIME type of the file.
  2. If file content sniffers do not identify the MIME type, then the application can check the filename. The application checks the filename against the MIME type registry. The MIME type registry associates particular file extensions and filename patterns, with particular MIME types. If a match for the filename is found, the MIME type associated with the extension or pattern is the MIME type of the file.

The following sections provide further information on file content sniffers and the MIME type registry.

5.2.1. File Content Sniffers

File content sniffers are specified in the file /etc/gnome-vfs-mime-magic. The following is an example of a file content sniffer:

0 string \x89PNG image/png

The syntax for file content sniffers is as follows:

offset_start[:offset_end] pattern_type pattern [&pattern_mask] type

Table 5-1 describes the fields in a file content sniffer.

Table 5-1Fields in a File Content Sniffer
Field Description
offset_start Specifies the number of characters to ignore in the file before searching for a text pattern.
pattern_type Specifies the type of pattern to search for. The string pattern type is the only pattern type that is supported at the time of publication of this guide.
pattern Specifies the pattern to search for.
pattern_mask

Specifies a pattern mask, in hexadecimal format. For more information on pattern masks, see the next section.

This field is optional. This field is not present in the example.

type Specifies the MIME type to associate with files that match this entry.

5.2.2. Pattern Masks

A pattern mask identifies bits in the pattern to ignore when searching for a pattern in a file. The following is an example of a file content sniffer with a pattern mask:

0 string BMxxxx\000\000 &0xffff00000000ffff image/bmp

The pattern and mask in the example are as follows:

Pattern B M x x x x \000 \000
Mask ff ff 00 00 00 00 ff ff

The pattern and mask specify a file with the following characteristics:

  1. The file begins with BM.
  2. BM is followed by four bytes with any values.
  3. The four bytes are followed by \000\000.

The file content sniffer specifies that the MIME type of files that match the pattern and mask is image/bmp.

5.2.3. MIME Type Registry

The MIME type registry is located in /usr/share/mime-info. The MIME type registry contains the following files:

File File Extension
MIME information file .mime
MIME keys file .keys

The following sections describe MIME information files and MIME keys files.

5.2.3.1. MIME Information Files

MIME information files associate MIME types with one or both of the following:

  • File extensions
  • Filename patterns

When an application searches for the MIME type of a file, the application checks the filename against the MIME information files. If a match for the filename is found, the MIME type associated with the extension or pattern is the MIME type of the file.

In MIME information files, the filename pattern to search for is written as a regular expression.

The format of MIME type entries in MIME information files is as follows:

MIME-type ext[,priority]: list-of-extensions regex[,priority]: list-of-regular-expressions

You can specify a priority value for the file extension and the regular expression. You can use the priority value to differentiate composite filenames. For example, you can assign a priority of 1 to the .gz extension, and assign a higher priority of 2 to the .tar.gz extension. In this case, the file abc.tar.gz takes the MIME type for .tar.gz.

You must indent the ext field and the regex field with a tab character (\t).

The following MIME type entries are samples from the gnome-vfs.mime MIME information file:

application/x-compressed-tar regex,2: tar\.gz$ ext: tgz audio/x-real-audio ext: rm ra ram image/jpeg ext: jpe jpeg jpg image/png ext: png text/html ext: html htm HTML text/plain ext: asc txt TXT text/x-readme regex: README.*

The file manager reads the MIME information files alphabetically. The alphabetical order determines the order in which MIME types are assigned to file extensions or regular expressions. For example, if the same file extension is assigned to different MIME types in the files abc.mime and def.mime, the MIME type in abc.mime is used.

5.2.3.2. MIME Keys Files

MIME keys file provide information about a MIME type that is used in the user interface. For example, the MIME keys file provides a description of a MIME type, and specifies an icon to represent files of that MIME type.

The following is a sample from a MIME keys file:

text/html description=HTML page icon_filename=gnome-text-html default_action_type=application short_list_application_ids_for_novice_user_level=mozilla,netscape,galeon category=Documents/World Wide Web

You must indent the keys in a MIME keys file with a tab character (\t).

Table 5-2 describes the most important keys in MIME keys files. Typically, the description key and the category key are localized.

Table 5-2Keys in MIME Keys Files
Key Description
can_be_executable Specifies whether files of this MIME type can be executed.
description Describes the MIME type. This description can be displayed in the file manager and other applications.
icon_filename

Specifies the filename of an icon to represent the MIME type. Does not specify the path to the filename, or the file extension.

This icon can be displayed in the file manager and other applications.

default_action_type Specifies the category of action to take when a file of this MIME type is opened by the user. Enter application for this MIME type for most applications.
short_list_application_ids _for_novice_user_level
Specifies the application to use when a file of this MIME type is opened by a user. Specify one or more applications, in order of priority. The applications must also be registered in the application registry.
category Specifies a category for the MIME type. The value of this key determines the location of the MIME type in the File Types and Programs preference tool.