diff -uNr Mail-SpamAssassin-3.2.3.orig/lib/Mail/SpamAssassin/HTML.pm Mail-SpamAssassin-3.2.3/lib/Mail/SpamAssassin/HTML.pm
--- Mail-SpamAssassin-3.2.3.orig/lib/Mail/SpamAssassin/HTML.pm	2007-08-08 22:19:15.000000000 +0900
+++ Mail-SpamAssassin-3.2.3/lib/Mail/SpamAssassin/HTML.pm	2007-12-27 15:59:40.000000000 +0900
@@ -86,7 +86,9 @@
 $ok_attributes{span}{$_} = 1 for qw( style );
 
 sub new {
-  my ($class) = @_;
+  my $class = shift;
+  my $opts = shift;
+
   my $self = $class->SUPER::new(
 		api_version => 3,
 		handlers => [
@@ -99,6 +101,7 @@
 			declaration => ["html_declaration", "self,text"],
 		],
 		marked_sections => 1);
+  $self->{normalize} = $opts->{normalize} || 0;
 
   $self;
 }
@@ -680,7 +683,14 @@
     }
   }
   else {
-    $text =~ s/[ \t\n\r\f\x0b\xa0]+/ /g;
+    if ($self->{normalize}) {
+      $text =~ s/\xc2\xa0/ /g;           # no-break space
+      $text =~ s/\xe3\x80\x80/ /g;       # ideographicspace
+      $text =~ s/[ \t\n\r\f\x0b]+/ /g;
+    }
+    else {
+      $text =~ s/[ \t\n\r\f\x0b\xa0]+/ /g;
+    }
     # trim leading whitespace if previous element was whitespace
     if (@{ $self->{text} } &&
 	defined $self->{text_whitespace} &&
diff -uNr Mail-SpamAssassin-3.2.3.orig/lib/Mail/SpamAssassin/Message/Node.pm Mail-SpamAssassin-3.2.3/lib/Mail/SpamAssassin/Message/Node.pm
--- Mail-SpamAssassin-3.2.3.orig/lib/Mail/SpamAssassin/Message/Node.pm	2007-08-08 22:19:11.000000000 +0900
+++ Mail-SpamAssassin-3.2.3/lib/Mail/SpamAssassin/Message/Node.pm	2007-12-27 15:59:40.000000000 +0900
@@ -41,6 +41,7 @@
 use Mail::SpamAssassin::Constants qw(:sa);
 use Mail::SpamAssassin::HTML;
 use Mail::SpamAssassin::Logger;
+use Mail::SpamAssassin::Util::Charset;
 
 =item new()
 
@@ -378,27 +379,10 @@
 
 sub _normalize {
   my ($self, $data, $charset) = @_;
-  return $data unless $self->{normalize};
+  return wantarray ? ($data, $charset) : $data unless $self->{normalize};
 
-  my $detected = Encode::Detect::Detector::detect($data);
-
-  my $converter;
-
-  if ($charset && $charset !~ /^us-ascii$/i &&
-      ($detected || 'none') !~ /^(?:UTF|EUC|ISO-2022|Shift_JIS|Big5|GB)/i) {
-      dbg("Using labeled charset $charset");
-      $converter = Encode::find_encoding($charset);
-  }
-
-  $converter = Encode::find_encoding($detected) unless $converter || !defined($detected);
-
-  return $data unless $converter;
-
-  dbg("Converting...");
-
-  my $rv = $converter->decode($data, 0);
-  utf8::downgrade($rv, 1);
-  return $rv
+  my ($decoded_data, $detected_charset) = normalize_charset($data, $charset);
+  return wantarray ? ($decoded_data, $detected_charset) : $decoded_data;
 }
 
 =item rendered()
@@ -421,8 +405,12 @@
     # text/x-aol is ignored here, but looks like text/html ...
     return(undef,undef) unless ( $self->{'type'} =~ /^text\/(?:plain|html)$/i );
 
-    my $text = $self->_normalize($self->decode(), $self->{charset});
+    my ($text, $charset) = $self->_normalize($self->decode(), $self->{charset});
     my $raw = length($text);
+    if ($self->{normalize}) {
+        $self->{charset} = $charset;
+        $self->{language} = get_language($text, $charset);
+    }
 
     # render text/html always, or any other text|text/plain part as text/html
     # based on a heuristic which simulates a certain common mail client
@@ -432,7 +420,7 @@
     {
       $self->{rendered_type} = 'text/html';
 
-      my $html = Mail::SpamAssassin::HTML->new();	# object
+      my $html = Mail::SpamAssassin::HTML->new({normalize=>$self->{normalize}});	# object
       $html->parse($text);				# parse+render text
       $self->{rendered} = $html->get_rendered_text();
       $self->{visible_rendered} = $html->get_rendered_text(invisible => 0);
diff -uNr Mail-SpamAssassin-3.2.3.orig/lib/Mail/SpamAssassin/Message.pm Mail-SpamAssassin-3.2.3/lib/Mail/SpamAssassin/Message.pm
--- Mail-SpamAssassin-3.2.3.orig/lib/Mail/SpamAssassin/Message.pm	2007-08-08 22:19:15.000000000 +0900
+++ Mail-SpamAssassin-3.2.3/lib/Mail/SpamAssassin/Message.pm	2007-12-27 15:59:40.000000000 +0900
@@ -538,6 +538,8 @@
   delete $self->{'pristine_headers'};
   delete $self->{'line_ending'};
   delete $self->{'missing_head_body_separator'};
+  delete $self->{'charset'};
+  delete $self->{'language'};
 
   my @toclean = ( $self );
 
@@ -553,6 +555,8 @@
     delete $part->{'invisible_rendered'};
     delete $part->{'type'};
     delete $part->{'rendered_type'};
+    delete $self->{'charset'};
+    delete $self->{'language'};
 
     # if there are children nodes, add them to the queue of nodes to clean up
     if (exists $part->{'body_parts'}) {
@@ -933,7 +937,14 @@
 
   # whitespace handling (warning: small changes have large effects!)
   $text =~ s/\n+\s*\n+/\f/gs;		# double newlines => form feed
-  $text =~ tr/ \t\n\r\x0b\xa0/ /s;	# whitespace => space
+  if ($self->{normalize}) {
+    $text =~ s/\xc2\xa0/ /g;           # no-break space => space
+    $text =~ s/\xe3\x80\x80/ /g;       # ideographicspace => space
+    $text =~ tr/ \t\n\r\x0b/ /s;       # whitespace => space
+  }
+  else {
+    $text =~ tr/ \t\n\r\x0b\xa0/ /s;   # whitespace => space
+  }
   $text =~ tr/\f/\n/;			# form feeds => newline
   
   # warn "message: $text";
@@ -990,7 +1001,14 @@
 
   # whitespace handling (warning: small changes have large effects!)
   $text =~ s/\n+\s*\n+/\f/gs;		# double newlines => form feed
-  $text =~ tr/ \t\n\r\x0b\xa0/ /s;	# whitespace => space
+  if ($self->{normalize}) {
+    $text =~ s/\xc2\xa0/ /g;           # no-break space => space
+    $text =~ s/\xe3\x80\x80/ /g;       # ideographicspace => space
+    $text =~ tr/ \t\n\r\x0b/ /s;       # whitespace => space
+  }
+  else {
+    $text =~ tr/ \t\n\r\x0b\xa0/ /s;   # whitespace => space
+  }
   $text =~ tr/\f/\n/;			# form feeds => newline
 
   my @textary = split_into_array_of_short_lines ($text);
@@ -1041,7 +1059,14 @@
 
   # whitespace handling (warning: small changes have large effects!)
   $text =~ s/\n+\s*\n+/\f/gs;		# double newlines => form feed
-  $text =~ tr/ \t\n\r\x0b\xa0/ /s;	# whitespace => space
+  if ($self->{normalize}) {
+    $text =~ s/\xc2\xa0/ /g;           # no-break space => space
+    $text =~ s/\xe3\x80\x80/ /g;       # ideographicspace => space
+    $text =~ tr/ \t\n\r\x0b/ /s;       # whitespace => space
+  }
+  else {
+    $text =~ tr/ \t\n\r\x0b\xa0/ /s;   # whitespace => space
+  }
   $text =~ tr/\f/\n/;			# form feeds => newline
 
   my @textary = split_into_array_of_short_lines ($text);
@@ -1094,6 +1119,28 @@
 
 # ---------------------------------------------------------------------------
 
+sub get_language {
+  my ($self) = @_;
+
+  if (defined $self->{language}) { return $self->{language}; }
+  my @parts = $self->find_parts(qr/^(?:text|message)\b/i,1);
+  return '' unless @parts;
+
+  # Go through each part
+  my @langs;
+  for(my $pt = 0 ; $pt <= $#parts ; $pt++ ) {
+    my $p = $parts[$pt];
+    my $lang = $p->{language};
+    next unless ($lang);
+    push(@langs, $lang) unless (grep(/^$lang$/, @langs))
+  }
+  $self->{language} = scalar(@langs) ? join(' ', @langs) : '';
+  return $self->{language};
+}
+
+# ---------------------------------------------------------------------------
+
+
 1;
 
 =back