File::Type::Builder(3pm)

File::Type::Builder3pm

Parse mime-magic and generate code

Usage

    my $build = File::Type::Builder->new();
    
    while (<magic>) {
      chomp;
      my $parsed = $build->parse_magic($_);
      
      my $code   = $build->string_start($parsed);
      (or string_offset or beshort)
    }

Overview

Reads in the mime-magic file format and translates it to code. (This documentation would be longer if I really expected anyone other than me to run the code.)

Methods

new

Creates a new File::Type::Builder object.

parse_magic

Pulls apart a line of a mime-magic file using a string of regular expressions.

An example mime-magic file can be found in cleancode CVS at <http://cleancode.org/cgi-bin/viewcvs.cgi/email/mime-magic.mime?rev=1.1.1.1>

string

Builds code to match magic that's of type string.

Has to do some cleverness to make the regular expression work properly.

be

Builds code to match 'beshort' and 'belong' magic (eg audio/mpeg, image/jpeg).

Private methods

_substr_matching

Sometimes the data is smaller than the offset we're looking for in the file. If this is the case, then the file is obviously not of that type, and furthermore we should avoid issuing a couple of warnings that Perl would otherwise emit.

This subroutine generates this code.

_get_escapes

Returns a reference to a hash defining characters that should not be escaped.

Todo

* Add handlers for other magic types (bedate, byte, etc)

* Make verbosity/logging nicer.

* Find more edge cases.

* Remove redundant 'if (length $data > 0)' check.

Longer term:

* Fix for multiple magic format types?

Releases

12 releases since Debian 3.1 (Sarge); the text changed 6 times.

  1. 3.1–4.0
  2. 5.0–7
  3. 8
  4. 9–10
  5. 11
  6. 12–13
  7. Arch

Bugs

Incomplete. Some known issues with odd entries in mime-magic. Skips some mime-magic lines.

Author

Paul Mison <[email protected]>

License

Licensed under the same terms as Perl itself.

From perl-file-type 0.22-18 in Arch Linux 2026-11-01. Parsed by carta from the page's man source.