view release on metacpan or  search on metacpan

lib/Mail/SpamAssassin/Pyzor/Digest/Pieces.pm  view on Meta::CPAN

sub parse_content_type {
    my ($content_type) = @_;

    # text/plain; charset=us-ascii
    my $ct_parse;
    if($content_type =~ /(\w+)\/(\w+); charset=(.*)/) {
      $ct_parse->{type} = $1;
      $ct_parse->{subtype} = $2;
      $ct_parse->{'attributes'}{'charset'} = $3;
    } elsif($content_type =~ /(\w+)\/(\w+)/) {
      $ct_parse->{type} = $1;
      $ct_parse->{subtype} = $2;
      $ct_parse->{'attributes'}{'charset'} = 'us-ascii';
    } else {
      $ct_parse->{type} = 'text';
      $ct_parse->{subtype} = 'plain';
      $ct_parse->{'attributes'}{'charset'} = 'us-ascii';
    }

    my $main = $ct_parse->{'type'}    || q<>;
    my $sub  = $ct_parse->{'subtype'} || q<>;

    my $encoding = $ct_parse->{'attributes'}{'charset'};

    my $checkval;

    if ($encoding) {

        # Lower-case everything, convert underscore to dash, and remove NUL.
        $encoding =~ tr<A-Z_\0><a-z->d;

        # Apparently pyzor accommodates messages that put the transfer
        # encoding in the Content-Type.
        if ( grep { $_ eq $encoding } _QUOTED_PRINTABLE_NAMES() ) {
            $checkval = Encode::FB_CROAK();
        }
    }
    else {
        $encoding = 'ascii';
    }

    # Match Python .decode()'s 'ignore' behavior
    $checkval ||= \&_encode_check_ignore;

    return ( $main, $sub, $encoding, $checkval );
}

#----------------------------------------------------------------------

=head2 @lines = splitlines( $TEXT )

Imitates C<str.splitlines()>. (cf. C<pydoc str>)

Returns a plain list in list context. Returns the number of
items to be returned in scalar context.

=cut

sub splitlines {
    return split m<\r\n?|\n>, $_[0] if defined $_[0];
}

1;