find_encoding returns Internal encoding

[Rikkuru]

Hi I caught an error " Unknown encoding 'Internal' " in call to from_to with encoding object returned from find_encoding.
Here is an example:

perl -e 'use strict; use warnings; use Encode; my $enc = Encode::find_encoding("Unicode"); Encode::from_to("abc", $enc->name, "utf-8");'

I think Internal encoding obj should not be returned if it cant be used.

[pali]

It is caused by following code, which defines Unicode encoding via Internal package.

p5-encode/Encode.pm

Lines 176 to 217 in 67f59dd

	if ($ON_EBCDIC) {
	package Encode::UTF_EBCDIC;
	use parent 'Encode::Encoding';
	my $obj = bless { Name => "UTF_EBCDIC" } => "Encode::UTF_EBCDIC";
	Encode::define_encoding($obj, 'Unicode');
	sub decode {
	my ( undef, $str, $chk ) = @_;
	my $res = '';
	for ( my $i = 0 ; $i < length($str) ; $i++ ) {
	$res .=
	chr(
	utf8::unicode_to_native( ord( substr( $str, $i, 1 ) ) )
	);
	}
	$_[1] = '' if $chk;
	return $res;
	}
	sub encode {
	my ( undef, $str, $chk ) = @_;
	my $res = '';
	for ( my $i = 0 ; $i < length($str) ; $i++ ) {
	$res .=
	chr(
	utf8::native_to_unicode( ord( substr( $str, $i, 1 ) ) )
	);
	}
	$_[1] = '' if $chk;
	return $res;
	}
	} else {
	package Encode::Internal;
	use parent 'Encode::Encoding';
	my $obj = bless { Name => "Internal" } => "Encode::Internal";
	Encode::define_encoding($obj, 'Unicode');
	sub decode {
	my ( undef, $str, $chk ) = @_;
	utf8::upgrade($str);
	$_[1] = '' if $chk;
	return $str;
	}
	*encode = \&decode;
	}

And I'm not sure what if purpose of that code. Because it contains following implementation of encode and decode methods:

p5-encode/Encode.pm

Lines 210 to 216 in 67f59dd

	sub decode {
	my ( undef, $str, $chk ) = @_;
	utf8::upgrade($str);
	$_[1] = '' if $chk;
	return $str;
	}
	*encode = \&decode;

@dankogai would probably know more why it is needed...

[dankogai]

Even I cannot recall what this code was for but since it is w/ UTF_EBCDIC this code existed BEFORE I took over the maintainer from late Nick Ing-Simmons. Luckily virtually no one used Unicode as encoding names till this report. Since Unicode is not a name of encoding, this code should be removed.

[dankogai]

Done. 06591f7
Unicode is no longer a valid encoding name, as it should be.