Author image Péter Szabó


MIME::AltWords - properly deal with RFC-1522 encoded words


The Perl module MIME::AltWords is recommended for encoding and decoding MIME words (such as =?ISO-8859-2?Q?_=E1ll_e=E1r?=) found in e-mail message headers (mostly Subject, From and To).

MIME::AltWords is similar to MIME::Words in MIME::Tools, but it provides an alternate implementation that follows the MIME specification more carefully, and it is actually compatible with existing mail software (tested with Mutt, Pine, JavaMail and OpenWebmail). MIME::AltWords extends the functionality of MIME::Words (version 5.420) by adding more functions and more options to existing functions. The original interface is changed in an upward-compatible way.

Before reading further, you should see MIME::Tools to make sure that you understand where this module fits into the grand scheme of things. Go on, do it now. I'll wait.

Ready? Ok...

    use MIME::AltWords qw(:all);   
    ### Decode the string into another string, forgetting the charsets:
    $decoded = decode_mimewords(
          'To: =?ISO-8859-1?Q?Keld_J=F8rn_Simonsen?= <>',
    ### Split string into array of decoded [DATA,CHARSET] pairs:
    @decoded = decode_mimewords(
          'To: =?ISO-8859-1?Q?Keld_J=F8rn_Simonsen?= <>',
    ### Encode a single unsafe word:
    $encoded = encode_mimeword("\xABFran\xE7ois\xBB");
    ### Encode a string, trying to find the unsafe words inside it: 
    $encoded = encode_mimewords("Me and \xABFran\xE7ois\xBB in town");


Fellow Americans, you probably won't know what the hell this module is for. Europeans, Russians, et al, you probably do. :-).

For example, here's a valid MIME header you might get:

      From: =?US-ASCII?Q?Keith_Moore?= <>
      To: =?ISO-8859-1?Q?Keld_J=F8rn_Simonsen?= <>
      CC: =?ISO-8859-1?Q?Andr=E9_?= Pirard <>
      Subject: =?ISO-8859-1?B?SWYgeW91IGNhbiByZWFkIHRoaXMgeW8=?=

The fields basically decode to (sorry, I can only approximate the Latin characters with 7 bit sequences /o and 'e):

      From: Keith Moore <>
      To: Keld J/orn Simonsen <>
      CC: Andr'e  Pirard <>
      Subject: If you can read this you understand the example... cool!


encode_mimewords RAW, [OPTS]

Function. Given a RAW string, try to find and encode all "unsafe" sequences of characters:

    ### Encode a string with some unsafe "words":
    $encoded = encode_mimewords("Me and \xABFran\xE7ois\xBB");

Returns the encoded string. Any arguments past the RAW string are taken to define a hash of options:


Encode all unsafe stuff with this charset. Default is 'ISO-8859-1', a.k.a. "Latin-1".


The encoding to use, "q" or "b". The default is "q".


Name of the mail field this string will be used in. Currently ignored.

Note: this is a stable, tested, widely compatible solution. Strict compliance with RFC-1522 (regarding the use of encoded words in message headers), however, was not proven, but strings returned by this function work properly and identically with Mutt, Pine, JavaMail and OpenWebmail. The recommended way is to use this function instead of encode_mimeword() or "encode_mimewords" in MIME::Words.

encode_mimeword RAW, [ENCODING], [CHARSET]

Function. Encode a single RAW "word" that has unsafe characters. The "word" will be encoded in its entirety.

    ### Encode "<<Franc,ois>>":
    $encoded = encode_mimeword("\xABFran\xE7ois\xBB");

You may specify the ENCODING ("Q" or "B"), which defaults to "Q". You may specify the CHARSET, which defaults to iso-8859-1.

decode_mimewords ENCODED, [OPTS...]

Function. Go through the string looking for RFC-1522-style "Q" (quoted-printable, sort of) or "B" (base64) encoding, and decode them.

In an array context, splits the ENCODED string into a list of decoded [DATA, CHARSET] pairs, and returns that list. Unencoded data are returned in a 1-element array [DATA], giving an effective CHARSET of undef.

    $enc = '=?ISO-8859-1?Q?Keld_J=F8rn_Simonsen?= <>';
    foreach (decode_mimewords($enc)) {
        print "", ($_[1] || 'US-ASCII'), ": ", $_[0], "\n";

In a scalar context, joins the "data" elements of the above list together, and returns that. Note: this is not information-lossy, it sanitizes the returned string to use a specific, single charset, either specified using the Charset option, or autodetecting one (ISO-8859-1, ISO-8859-2 or UTF-8) which can accomodate all characters. In case of charset autodetection, get_best_decode_charset(ENCODED) can be used to query the charset autodetected.

You might want to see "unmime" in MIME::WordDecoder as an alternate of MIME::AltWords::encode_mimewords.

In the event of a syntax error, $@ will be set to a description of the error, but parsing will continue as best as possible (so as to get something back when decoding headers). $@ will be false if no error was detected.

Any arguments past the ENCODED string are taken to define a hash of options:


Name of the mail field this string came from. Currently ignored.


Exports its principle functions by default, in keeping with MIME::Base64 and MIME::QuotedPrint.

Doesn't depend on MIME::Words or MIME::Tools.

See also for the previous version of MIME::AltWords integrated into the Sympa 4 mailing list software.


MIME::AltWords was written by Péter Szabó ( in 2006, and it has been uploaded to CPAN on 2006-09-27.

MIME::AltWords uses code from MIME::Words (in the function decode_mimewords_wantarray) and it uses documentation from MIME::Words (in the file lib/MIME/

Here is the original author and copyright information for MIME::Words.

Eryq (, ZeeGee Software Inc ( David F. Skoll (

All rights reserved. This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

Thanks also to...

      Kent Boortz        For providing the idea, and the baseline 
                         RFC-1522-decoding code!
      KJJ at PrimeNet    For requesting that this be split into
                         its own module.
      Stephane Barizien  For reporting a nasty bug.


See $VERSION in lib/MIME/ .