Fix bug in conversion process between RSS and Atom where HTML got double encoded

[catagits/XML-Feed.git] / lib / XML / Feed.pm

diff --git a/lib/XML/Feed.pm b/lib/XML/Feed.pm
index 6a0ce28..7f609f3 100644 (file)
--- a/lib/XML/Feed.pm
+++ b/lib/XML/Feed.pm
@@ -1,4 +1,4 @@
-# $Id: Feed.pm 1958 2006-08-14 05:31:27Z btrott $
+# $Id$
 
 package XML::Feed;
 use strict;
@@ -8,16 +8,24 @@ use Feed::Find;
 use URI::Fetch;
 use LWP::UserAgent;
 use Carp;
-
-our $VERSION = '0.23';
+use Module::Pluggable search_path => "XML::Feed::Format",
+                      require     => 1,
+                      sub_name    => 'formatters';
+
+our $VERSION = '0.44';
+our $MULTIPLE_ENCLOSURES = 0;
+our @formatters;
+BEGIN {
+       @formatters = __PACKAGE__->formatters;
+}
 
 sub new {
     my $class = shift;
     my $format = shift || 'Atom';
-    my $format_class = 'XML::Feed::' . $format;
+    my $format_class = 'XML::Feed::Format::' . $format;
     eval "use $format_class";
     Carp::croak("Unsupported format $format: $@") if $@;
-    my $feed = bless {}, join('::', __PACKAGE__, $format);
+    my $feed = bless {}, join('::', __PACKAGE__, "Format", $format);
     $feed->init_empty(@_) or return $class->error($feed->errstr);
     $feed;
 }
@@ -61,7 +69,7 @@ sub parse {
         $format = $feed->identify_format(\$xml) or return $class->error($feed->errstr);
     }
 
-    my $format_class = join '::', __PACKAGE__, $format;
+    my $format_class = join '::', __PACKAGE__, "Format", $format;
     eval "use $format_class";
     return $class->error("Unsupported format $format: $@") if $@;
     bless $feed, $format_class;
@@ -70,8 +78,23 @@ sub parse {
 }
 
 sub identify_format {
-    my $feed = shift;
-    my($xml) = @_;
+    my $feed   = shift;
+    my($xml)   = @_;
+       foreach my $class (@formatters) {
+               my ($name) = ($class =~ m!([^:]+)$!);
+               # TODO ugly
+               my $tmp = $$xml;
+               return $name if eval { $class->identify(\$tmp) };
+               return $feed->error($@) if $@;
+       } 
+       return $feed->error("Cannot detect feed type");
+}
+
+sub _get_first_tag {
+       my $class  = shift;
+       my ($xml)  = @_;
+
+
     ## Auto-detect feed type based on first element. This is prone
     ## to breakage, but then again we don't want to parse the whole
     ## feed ourselves.
@@ -81,15 +104,9 @@ sub identify_format {
         my $first = substr $t, 0, 1;
         $tag = $t, last unless $first eq '?' || $first eq '!';
     }
-    return $feed->error("Cannot find first element") unless $tag;
+       die ("Cannot find first element") unless $tag;
     $tag =~ s/^.*://;
-    if ($tag eq 'rss' || $tag eq 'RDF') {
-        return 'RSS';
-    } elsif ($tag eq 'feed') {
-        return 'Atom';
-    } else {
-        return $feed->error("Cannot detect feed type");
-    }
+       return $tag;
 }
 
 sub find_feeds {
@@ -103,7 +120,7 @@ sub find_feeds {
 sub convert {
     my $feed = shift;
     my($format) = @_;
-    my $new = __PACKAGE__->new($format);
+    my $new = XML::Feed->new($format);
     for my $field (qw( title link description language author copyright modified generator )) {
         my $val = $feed->$field();
         next unless defined $val;
@@ -127,8 +144,8 @@ sub splice {
 sub _convert_entry {
     my $feed   = shift;
     my $entry  = shift;
-    my $feed_format  = ref($feed);   $feed_format  =~ s!^XML::Feed::!!;
-    my $entry_format = ref($entry);  $entry_format =~ s!^XML::Feed::Entry::!!;
+    my $feed_format  = ref($feed);   $feed_format  =~ s!^XML::Feed::Format::!!;
+    my $entry_format = ref($entry);  $entry_format =~ s!^XML::Feed::Entry::Format::!!;
     return $entry if $entry_format eq $feed_format;
     return $entry->convert($feed_format); 
 }
@@ -238,7 +255,7 @@ A URI from which the feed XML will be retrieved.
 
 =back
 
-C<$format> allows you to override format guessing.
+I<$format> allows you to override format guessing.
 
 =head2 XML::Feed->find_feeds($uri)
 
@@ -247,6 +264,10 @@ from that page (using I<E<lt>linkE<gt>> tags).
 
 Returns a list of feed URIs.
 
+=head2 XML::Feed->identify_format($xml)
+
+Given the xml of a feed return what format it is in (C<Atom>, or some version of C<RSS>).
+
 =head2 $feed->convert($format)
 
 Converts the I<XML::Feed> object into the I<$format> format, and returns
@@ -316,6 +337,10 @@ A string.
 A list of the entries/items in the feed. Returns an array containing
 I<XML::Feed::Entry> objects.
 
+=head2 $feed->items
+
+A synonym for I<$feed->entries>.
+
 =head2 $feed->add_entry($entry)
 
 Adds an entry to the feed. I<$entry> should be an I<XML::Feed::Entry>
@@ -340,8 +365,23 @@ package.
 
 B<Note:> this will only work for parsing feeds, not creating feeds.
 
+B<Note:> Only C<XML::RSS::LibXML> version 0.3004 is known to work at the moment.
+
+=item C<$XML::Feed::MULTIPLE_ENCLOSURES>
+
+Although the RSS specification states that there can be at most one enclosure per item 
+some feeds break this rule.
+
+If this variable is set then C<XML::Feed> captures all of them and makes them available as a list.
+
+Otherwise it returns the last enclosure parsed.
+
+B<Note:> C<XML::RSS> version 1.44 is needed for this to work.
+
 =back
 
+=cut
+
 =head1 VALID FEEDS
 
 For reference, this cgi script will create valid, albeit nonsensical feeds