Once the flag is on, does

use Data::Dumper;
$Data::Dumper::Useqq = 1;
print Dumper $string;
[download]

I'm not using an utf8 locale, nor does I use 'use utf8'.
My strings do not contain > 8 bit characters, but they contain 8 bit characters (German "Umlaute" - ä, ö, ü, ß).

I'm not using an utf8 locale, nor does I use 'use utf8'.

The comment about 'use utf8' here makes me think it's just possible you are mistaken. "Using a utf8 locale" isn't a perl thing, it's an operating system thing, and several OS's have taken to setting utf8 locales as a default. Check your environment variables LANG, LC_CALL, LC_CTYPE, and LANGUAGE.

If that's not it, you'll have to dig a little bit into what happens to this variable; at what point does it get the utf8 flag turned on?

  my $octets = encode( 'iso-8859-15', $utf8string );
[download]

I was using encode() correctly. I found my mistake - and it was a strange one!

I was using encode() in a recursive function (to convert my datastructure entirely to iso-8859-15).

sub recode {
    my ($enc, $data) = @_;

    my $ref = ref($data);
    if ($ref) {
        if ($ref eq 'ARRAY') {
            # Array-Ref
            my @array = map { recode($enc, $_) } @$data;
            return \@array;
        } elsif ($ref eq 'SCALAR') {
            # Scalar-Ref
            my $scalar = recode($enc, $data);
            return \$scalar;
        } elsif ($ref eq 'HASH') {
            # Hash-Ref
            my %hash = ();
            while (my ($key, $value) = each %$data) {
                $hash{recode($enc, $key)} = recode($enc, $value);
            }
            return \%hash;
        } else {
            # Object - XYZ::, ZYX::
            if ($ref =~ /^(XYZ|ZYX)::/) {
                my $object = bless({}, $ref);
                while (my ($key, $value) = each %$data) {
                    $object->{recode($enc, $key)} = recode($enc, $valu
+e);
                }
                return $object;
            } else {
                warn "recode(): $ref nicht unterstützt";
                return $data;
            }
        }
    } else {
        # unbedingt Variable verwenden, sonst wird
        # das UTF8-Flag nicht gelöscht
        $data = Encode::encode($enc, $data);

        return $data;
    }
}
[download]

(Please ignore the German comments.)

The problem was the last else-block. I was using "return Encode::encode($enc, $data);" without a variable assignment. And this code did not clear the utf8 flag! Only with a variable assignment the utf8 flag was cleared.

I don't know why this happens. I tried to reproduce this behaviour to report a bug, but I failed:

#!/usr/bin/perl -w

use strict;
use Devel::Peek;
use Encode ();

my $str = 'Übel';
Dump($str);

$str = Encode::decode('iso-8859-15', $str);
Dump($str);

my $str2 = do_encode($str);
Dump($str2);

my $str3 = do_encode_with_tmp($str);
Dump($str3);


sub do_encode {
    my $text = shift;

    return Encode::encode('iso-8859-15', $text);
}

sub do_encode_with_tmp {
    my $text = shift;

    my $tmp = Encode::encode('iso-8859-15', $text);

    return $tmp;
}
[download]

Here both subs clear the utf8 flag correctly....

Bye,
Uwe

v5.8.0 had some issues with its handling of UTF-8 that were fixed in 5.8.1. I would highly recommend upgrading, if possible. If not, here are some pointers to changes:

• From perl581delta: UTF-8 On Filehandles No Longer Activated By Locale
• From perl581delta: UTF-8 no longer default under UTF-8 locales

Also in the delta is this information about updates to the Encode module:

Significant updates on the encoding pragma functionality (tr/// and the DATA filehandle, formats).
If a filehandle has been marked as to have an encoding, unmappable characters are detected already during input, not later (when the corrupted data is being used).
...

Hope this helps