Encode::Newlines - Normalize line ending sequences


This document describes version 0.04 of Encode::Newlines, released September 4, 2007.


    use Encode;
    use Encode::Newlines;

    # Convert to native newlines
    # Note that decode() and encode() are equivalent here
    $native = decode(Native => $string);
    $native = encode(Native => $string);

        # Allow mixed newlines in $mixed
        local $Encode::Newlines::AllowMixed = 1;
        $cr = encode(CR => $mixed);


This module provides the CR, LF, CRLF and Native encodings, to aid in normalizing line endings.

It converts whatever line endings the source uses to the designated newline sequence, for both encode and decode operations.

If you specify two different line endings joined by a -, it will use the first one for decoding and the second one for encoding. For example, the LF-CRLF encoding means that all input should be normalized to LF, and all output should be normalized to CRLF.

If the source has an inconsistent line ending style, then a Mixed newlines exception is raised on behalf of the caller. However, if the package variable $Encode::Newlines::AllowMixed is set to a true value, then it will silently convert all three line endings.


This module is not suited for working with PerlIO::encoding, because it cannot guarantee that the chunk bounaries won't happen within a CR/LF sequence. See PerlIO::eol for how to deal with this correctly.

An optional XS implemenation would be nice.


Audrey Tang <>


Copyright 2004-2007 by Audrey Tang <>.

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