X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=Debbugs%2FMIME.pm;h=05534e3e054c47d2cb281f683aaacf1a0ac7a52e;hb=4ac94f292bfbf36863795ed9d695be7ad2b5b344;hp=95dafb8dc5075ed9a515143c00dd8e584e653bb0;hpb=c775dcf61c6a5e45303684d3cf5a3d27c3730698;p=debbugs.git

diff --git a/Debbugs/MIME.pm b/Debbugs/MIME.pm
index 95dafb8..05534e3 100644
--- a/Debbugs/MIME.pm
+++ b/Debbugs/MIME.pm
@@ -229,9 +229,7 @@ sub convert_to_utf8 {
      return $data if $charset eq 'raw' or is_utf8($data,1);
      my $result;
      eval {
-	  # this encode/decode madness is to make sure that the data
-	  # really is valid utf8 and that the is_utf8 flag is off.
-	  $result = encode("utf8",decode($charset,$data))
+	 $result = decode($charset,$data);
      };
      if ($@) {
 	  warn "Unable to decode charset; '$charset' and '$data': $@";
@@ -286,10 +284,15 @@ sub encode_rfc1522 {
 
      # handle being passed undef properly
      return undef if not defined $rawstr;
+     if (is_utf8($rawstr)) {
+	 $rawstr= encode_utf8($rawstr);
+     }
      # We process words in reverse so we can preserve spacing between
      # encoded words. This regex splits on word|nonword boundaries and
-     # nonword|nonword boundaries.
-     my @words = reverse split /(?:(?<=[\s\n])|(?=[\s\n]))/m, $rawstr;
+     # nonword|nonword boundaries. We also consider parenthesis and "
+     # to be nonwords to avoid escaping them in comments in violation
+     # of RFC1522
+     my @words = reverse split /(?:(?<=[\s\n\)\(\"])|(?=[\s\n\)\(\"]))/m, $rawstr;
 
      my $previous_word_encoded = 0;
      my $string = '';
@@ -311,7 +314,7 @@ sub encode_rfc1522 {
 	       if (length $encoded > 75) {
 		    # Turn utf8 into the internal perl representation
 		    # so . is a character, not a byte.
-		    my $tempstr = decode_utf8($word,Encode::FB_DEFAULT);
+		    my $tempstr = is_utf8($word)?$word:decode_utf8($word,Encode::FB_DEFAULT);
 		    my @encoded;
 		    # Strip it into 10 character long segments, and encode
 		    # the segments