move UTF8 routines out of Debbugs::Common and into Debbugs::UTF8 so Debbugs::Config...

[debbugs.git] / Debbugs / MIME.pm

diff --git a/Debbugs/MIME.pm b/Debbugs/MIME.pm
index 95dafb8dc5075ed9a515143c00dd8e584e653bb0..482a30b4d182e530bbcd3b780a5479dc53f4cae4 100644 (file)
--- a/Debbugs/MIME.pm
+++ b/Debbugs/MIME.pm
@@ -31,12 +31,20 @@ use warnings;
 use strict;
 
 use base qw(Exporter);
-use vars qw($VERSION @EXPORT_OK);
+use vars qw($DEBUG $VERSION @EXPORT_OK %EXPORT_TAGS @EXPORT);
 
 BEGIN {
     $VERSION = 1.00;
+    $DEBUG = 0 unless defined $DEBUG;
 
-    @EXPORT_OK = qw(parse decode_rfc1522 encode_rfc1522 convert_to_utf8 create_mime_message getmailbody);
+    @EXPORT = ();
+
+    %EXPORT_TAGS = (mime => [qw(parse create_mime_message getmailbody)],
+                   rfc1522 => [qw(decode_rfc1522 encode_rfc1522)],
+                  );
+    @EXPORT_OK=();
+    Exporter::export_ok_tags(keys %EXPORT_TAGS);
+    $EXPORT_TAGS{all} = [@EXPORT_OK];
 }
 
 use File::Path;
@@ -46,6 +54,9 @@ use MIME::Parser;
 use POSIX qw(strftime);
 use List::MoreUtils qw(apply);
 
+# for convert_to_utf8
+use Debbugs::UTF8 qw(convert_to_utf8);
+
 # for decode_rfc1522
 use MIME::WordDecoder qw();
 use Encode qw(decode encode encode_utf8 decode_utf8 is_utf8);
@@ -60,7 +71,7 @@ sub getmailbody
     if ($type eq 'text/plain' or
            ($type =~ m#text/?# and $type ne 'text/html') or
            $type eq 'application/pgp') {
-       return $entity->bodyhandle;
+       return $entity;
     } elsif ($type eq 'multipart/alternative') {
        # RFC 2046 says we should use the last part we recognize.
        for my $part (reverse $entity->parts) {
@@ -92,8 +103,15 @@ sub parse
        @headerlines = @{$entity->head->header};
        chomp @headerlines;
 
-       my $entity_body = getmailbody($entity);
-       @bodylines = $entity_body ? $entity_body->as_lines() : ();
+        my $entity_body = getmailbody($entity);
+       my $entity_body_handle;
+        my $charset;
+        if (defined $entity_body) {
+            $entity_body_handle = $entity_body->bodyhandle();
+            $charset = $entity_body->head()->mime_attr('content-type.charset');
+        }
+       @bodylines = $entity_body_handle ? $entity_body_handle->as_lines() : ();
+        @bodylines = map {convert_to_utf8($_,$charset)} @bodylines;
        chomp @bodylines;
     } else {
        # Legacy pre-MIME code, kept around in case MIME::Parser fails.
@@ -185,7 +203,7 @@ sub create_mime_message{
      my $msg = MIME::Entity->build('Content-Type' => 'text/plain; charset=utf-8',
                                   'Encoding'     => 'quoted-printable',
                                   (map{encode_rfc1522($_)} @{$headers}),
-                                  Data    => $body
+                                  Data    => is_utf8($body)?encode_utf8($body):$body,
                                  );
 
      # Attach the attachments
@@ -220,25 +238,6 @@ sub create_mime_message{
 }
 
 
-# Bug #61342 et al.
-
-sub convert_to_utf8 {
-     my ($data, $charset) = @_;
-     # raw data just gets returned (that's the charset WordDecorder
-     # uses when it doesn't know what to do)
-     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,$data))
-     };
-     if ($@) {
-         warn "Unable to decode charset; '$charset' and '$data': $@";
-         return $data;
-     }
-     return $result;
-}
 
 
 =head2 decode_rfc1522
@@ -286,10 +285,15 @@ 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 +315,7 @@ 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