Perl error related to Encode.pm

Question

I have the following software:

Ubuntu Linux 14.04 LTS

$ uname -a
Linux XXX 3.13.0-45-generic #74-Ubuntu SMP Tue Jan 13 19:36:28 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

Perl 5.18:

$ perl -version

This is perl 5, version 18, subversion 2 (v5.18.2) built for x86_64-linux-gnu-thread-multi
(with 41 registered patches, see perl -V for more detail)

and Encode.pm 2.49:

$ head -n 10 /usr/lib/perl/5.18/Encode.pm
#
# $Id: Encode.pm,v 2.49 2013/03/05 03:13:47 dankogai Exp dankogai $

When I use ikiwiki, which is a package using Encode.pm, I got the following error:

$ ikiwiki --setup ~/wiki.setup
Cannot decode string with wide characters at /usr/lib/perl/5.18/Encode.pm line 176.

Lines 166 - 180 of Encode.pm reads:

sub decode($$;$) {
    my ( $name, $octets, $check ) = @_;
    return undef unless defined $octets;
    $octets .= '';
    $check ||= 0;
    my $enc = find_encoding($name);
    unless ( defined $enc ) {
        require Carp;
        Carp::croak("Unknown encoding '$name'");
    }
    my $string = $enc->decode( $octets, $check );
    $_[1] = $octets if $check and !ref $check and !( $check & LEAVE_SRC() );
    return $string;
}

Does any kind soul have idea how to fix this issue?

Answer 1

That error occurs when you try to decode something that's already decoded.

To expand a little, in Perl you can have byte strings and character strings. On input your data can be converted from bytes to a character string (e.g. using decode, or an IO layer). On output the character string must be converted to bytes (e.g. using encode or an IO layer).

It seems like in your case the $octets variable contains a string that has already been converted from bytes to characters. You could workaround this by adding an early return:

return $octets if Encode::is_utf8($octets);