Convert::YText3pm
Quotes strings suitably for rfc2822 local part
Usage
use Convert::YText qw(encode_ytext decode_ytext);
$encoded=encode_ytext($string); $decoded=decode_ytext($encoded);
($decoded eq $string) || die "this should never happen!";
Version
Version 0.2
Overview
Convert::YText converts strings to and from "YText", a format inspired by xtext defined in RFC1894, the MIME base64 and quoted-printable types (RFC 1394). The main goal is encode a UTF8 string into something safe for use as the local part in an internet email address (RFC2822).
By default spaces are replaced with "+", "/" with "~", the characters "A-Za-z0-9_.-" encode as themselves, and everything else is written "=USTR=" where USTR is the base64 (using "A-Za-z0-9_." as digits) encoding of the unicode character code. The encoding is configurable (see below).
Procedural interface
The module can can export "encode_ytext" which converts arbitrary unicode string into a "safe" form, and "decode_ytext" which recovers the original text. "validate_ytext" is a heuristic which returns 0 for bad input.
Object oriented interface.
For more control, you will need to use the OO interface.
new
Create a new encoding object.
Arguments
Arguments are by name (i.e. a hash).
- DIGIT_STRING ("A-Za-z0-9_.") Must be 64 characters long
- ESCAPE_CHAR ('=') Must not be in digit string.
- SPACE_CHAR ('+') Non digit to replace space. Can be the empty string.
- SLASH_CHAR ( '~') Non digit to replace slash. Can be the empty string.
- EXTRA_CHARS ('._\-') Other characters to leave unencoded.
encode
Arguments
a string to encode.
Returns
encoded string
decode
Arguments
a string to decode.
Returns
encoded string
valid
Simple necessary but not sufficient test for validity.
Discussion
According to RFC 2822, the following non-alphanumerics are OK for the local part of an address: "!#$%&'*+-/=?^_`{|}~". On the other hand, it seems common in practice to block addresses having "%!/|`#&?" in the local part. The idea is to restrict ourselves to basic ASCII alphanumerics, plus a small set of printable ASCII, namely "=_+-~.".
The characters '+' and '-' are pretty widely used to attach suffixes (although usually only one works on a given mail host). It seems ok to use '+-', since the first marks the beginning of a suffix, and then is a regular character. The character '.' also seems mostly permissible.
Releases
7 releases since Debian 7 (Wheezy); the text changed 4 times.