Hi,

I'm writing a system that will parse an email sent into a catchall email address. It's all working ok, but I'm having some fun with the base64 and QuotedPrint emails. Instead of trying to work it out myself (based on the enc type passed in), I've decided to opt for an existing module to work it out for me, and normalise it.

Here is a sample script I have:

    my @emails = split /\n/, q|foo <andy@bar.com>
=?utf-8?B?UGF1c2UgRG9yw6ll?= <pausedoree@gggg.com>
=?UTF-8?Q?Village_Bambous_=2D_Chambre_d=27_H=C3=B4tes?= <village.bambo
+us@ddd.com>
=?utf-8?B?YmVybmFyZCB2ZXJpdMOp?= <naturedetente@fdd.fr>
=?ISO-8859-1?B?TGHrdGl0aWE=?= Picot <villagabrielle@ffsdfsd.net>
=?iso-8859-1?Q?Ancie_chambres_d=27h=F4tes?= <ancie.ha@dfdd.fr>|;

    use utf8;
    use Encode qw(encode decode);

    foreach (@emails) {

            $_ = decode('MIME-Header', $_);
            print "FOO: $_\n";

            print $IN->header;
            use Data::Dumper;
            print Dumper($_);

            print "FOO: " . utf8::is_utf8($_) . "\n";

            if (utf8::is_utf8($_)) {
                print "content..\n";
                $_ =~ s/([\200-\377]+)/from_utf8({ -string => $1, -cha
+rset => 'ISO-8859-1'})/eg;
            }

            print Dumper($_);
            print "FOO: " . utf8::is_utf8($_) . "\n";

            print "\n\n";

    }
[download]

Here is a sample output:

FOO: foo <andy@bar.com>
Content-type: text/html; charset=iso-8859-1

$VAR1 = 'foo <andy@bar.com>';
FOO:
$VAR1 = 'foo <andy@bar.com>';
FOO:


FOO: Pause Dorée <pausedoree@gggg.com>
$VAR1 = "Pause Dor\x{e9}e <pausedoree\@gggg.com>";
FOO: 1
convert..
$VAR1 = "Pause Dor\x{e9}e <pausedoree\@gggg.com>";
FOO: 1


FOO: Village Bambous - Chambre d' Hôtes <village.bambous@ddd.com>
$VAR1 = "Village Bambous - Chambre d' H\x{f4}tes <village.bambous\@ddd
+.com>";
FOO: 1
convert..
$VAR1 = "Village Bambous - Chambre d' H\x{f4}tes <village.bambous\@ddd
+.com>";
FOO: 1


FOO: bernard verité <naturedetente@fdd.fr>
$VAR1 = "bernard verit\x{e9} <naturedetente\@fdd.fr>";
FOO: 1
convert..
$VAR1 = "bernard verit\x{e9} <naturedetente\@fdd.fr>";
FOO: 1


FOO: Laëtitia Picot <villagabrielle@ffsdfsd.net>
$VAR1 = "La\x{eb}titia Picot <villagabrielle\@ffsdfsd.net>";
FOO: 1
convert..
$VAR1 = "La\x{eb}titia Picot <villagabrielle\@ffsdfsd.net>";
FOO: 1


FOO: Ancie chambres d'hôtes <ancie.ha@dfdd.fr>
$VAR1 = "Ancie chambres d'h\x{f4}tes <ancie.ha\@dfdd.fr>";
FOO: 1
convert..
$VAR1 = "Ancie chambres d'h\x{f4}tes <ancie.ha\@dfdd.fr>";
FOO: 1
[download]

I'm a bit confused as to what encoding the string is in now though, as utf8::is_utf8($_) still seems to be giving me a positive, as to it being a utf8 string?

Basically, the end game is to have ANY encodings converted into iso-8859-1 format (I know I know, not ideal, but I'm dealing with a legacy system, and it would be months of work to convert the whole site into utf8)

Thanks for any suggestions

UPDATE: Interesting. The output of:

            $_ = decode('MIME-Header', $_);
[download]

Seems to actually give back the string in internal encoding. If I encode it after using:

    if (utf8::is_utf8($the_from)) {
        $the_from = encode('iso-8859-1', $_);
    }
[download]

That seems to do it. Does that look OK? I just don't want to bugger it up :)

Cheers

Andy