syndication 0.6 → 0.6.1

data/lib/syndication/common.rb

@@ -139,7 +139,8 @@ module Syndication
       'http://purl.org/rss/1.0/modules/content/' => 'content',
       'http://www.itunes.com/DTDs/Podcast-1.0.dtd' => 'itunes',
       'http://www.w3.org/1999/xhtml' => 'xhtml',
-      'http://schemas.google.com/g/2005' => 'gd'
+      'http://schemas.google.com/g/2005' => 'gd',
+      'http://rssnamespace.org/feedburner/ext/1.0' => 'feedburner'
     }
     
     # Create a new AbstractParser. The optional argument consists of text to
@@ -272,11 +273,16 @@ module Syndication
 
     # Supposed to be called when REXML finds a CDATA-encoded piece of text.
     def cdata(s)
-      # We re-encode, because (a) the API for RSS content module provides both 
-      # encoded and decoded results to the user, and (b) REXML doesn't always
-      # seem to pass CDATA via this callback method.
+      # For content_encoded we re-encode, because (a) the API for RSS content 
+      # module provides both encoded and decoded results to the user, and 
+      # (b) REXML doesn't always seem to pass CDATA via this callback method.
+      # For other elements, we keep the text decoded.
       if @textstack.last
-        @textstack.last << "<![CDATA[#{s}]]>"
+        if @tagstack.last == 'content:encoded'
+          @textstack.last << "<![CDATA[#{s}]]>"
+        else
+          @textstack.last << s
+        end
       end
     end
   end

data/lib/syndication/feedburner.rb

@@ -0,0 +1,18 @@
+
+module Syndication
+
+  module Feedburner
+    module Item
+      # The original URL, before feedburner rewrote it for tracking purposes
+      attr_accessor :feedburner_origlink
+    end
+
+  end
+
+  module RSS
+    class Item
+      include Feedburner::Item
+    end
+  end
+
+end

data/rakefile

@@ -5,7 +5,7 @@ require 'rake/gempackagetask'
 require 'rake/testtask'
 require 'rubygems'
 
-PKG_VERSION = "0.6"
+PKG_VERSION = "0.6.1"
 
 desc "Create HTML documentation from RDOC"
 Rake::RDocTask.new do |rd|
@@ -36,7 +36,7 @@ spec = Gem::Specification.new do |s|
   end
   s.require_path = "lib"
   s.test_files = ["test/atomtest.rb", "test/rsstest.rb", "test/google.rb",
-                  "test/tagsouptest.rb"]
+                  "test/tagsouptest.rb", "test/feedburntest.rb"]
   s.has_rdoc = true
   s.extra_rdoc_files = ["README", "IMPLEMENTATION", "CHANGES", "DEVELOPER"]
 end

data/test/feedburntest.rb

@@ -0,0 +1,79 @@
+# Copyright � mathew <meta@pobox.com> 2006.
+# Licensed under the same terms as Ruby.
+
+require 'syndication/rss'
+require 'test/unit'
+require 'syndication/dublincore'
+require 'syndication/content'
+require 'syndication/podcast'
+require 'syndication/feedburner'
+
+module Syndication
+
+# This class contains the unit tests for the Syndication module.
+class Tests < Test::Unit::TestCase
+
+    # A set of minimal assertions that can be applied to every well-formed parsed
+    # feed.
+    def baseline_rss_assertions(feed)
+      assert_not_nil(feed)
+      assert_kind_of(Syndication::RSS::Feed, feed)
+      loi = feed.items
+      assert_not_nil(loi)
+      assert_kind_of(Array, loi)
+      assert(loi.length >= 1)
+      assert_not_nil(loi[0])
+      assert_not_nil(loi[0].description)
+    end
+
+    def test_feedburner
+      xml = <<-EOF
+    <?xml version="1.0" encoding="UTF-8"?>
+    <?xml-stylesheet href="http://feeds.sfgate.com/~d/styles/rss2full.xsl" type="text/xsl" media="screen"?>
+    <?xml-stylesheet href="http://feeds.sfgate.com/~d/styles/itemcontent.css" type="text/css" media="screen"?>
+    <rss xmlns:feedburner="http://rssnamespace.org/feedburner/ext/1.0" version="2.0">
+    <channel>
+        <title>SFGate: Top News Stories</title>
+        <link>http://www.sfgate.com/</link>
+        <description>Top news stories. From SFGate.com: the Bay Area's home page, online home of the San Francisco Chronicle and much more.</description>
+        <language>en-us</language>
+        <copyright>Copyright 2006 Hearst Communications, Inc.</copyright>
+        <managingEditor>ed@sfgate.com (SFGate Editorial staff)</managingEditor>
+        <webMaster>support@sfgate.com (SFGate technical support)</webMaster>
+        <lastBuildDate>Sun, 09 Jul 2006 14:21:10 PDT</lastBuildDate>
+        <category>News</category>
+        <category>Newspapers</category>
+        <category>San Francisco</category>
+        <category>San Francisco Bay Area</category>
+        <docs>http://blogs.law.harvard.edu/tech/rss</docs>
+        <image>
+            <url>http://www.sfgate.com/templates/types/syndication/pages/rss/graphics/sfgate_logo.png</url>
+            <title>SFGate: Top News Stories</title>
+            <link>http://www.sfgate.com/</link>
+        </image>
+        <atom10:link xmlns:atom10="http://www.w3.org/2005/Atom" rel="self" href="http://www.sfgate.com/rss/feeds/news.xml" type="application/rss+xml" /><feedburner:browserFriendly>This is an RSS feed, but with the headlines made visible. Choose one of the buttons to add this feed to your favorite RSS reader.</feedburner:browserFriendly>
+        <item>
+            <title><![CDATA[Italy Beats France for 4th World Cup Title]]></title>
+            <link>http://feeds.sfgate.com/sfgate/rss/feeds/news?m=4300</link>
+            <description>Italy let France do nearly anything it wanted Sunday, except win the World Cup. That belongs to the Azzurri, 5-3 in a shootout after a 1-1 draw. Outplayed for an hour and into extra time, the Italians won it after French captain Zinedine Zidane was...&lt;img src="http://feeds.sfgate.com/sfgate/rss/feeds/news?g=4300"/&gt;</description>
+            <author><![CDATA[By BARRY WILNER, AP Sports Writer]]></author>
+            <pubDate>Sun, 09 Jul 2006 14:14:59 PDT</pubDate>
+
+            <guid isPermaLink="false">/n/a/2006/07/09/sports/s134544D82.DTL</guid>
+            <feedburner:origLink>http://www.sfgate.com/cgi-bin/article.cgi?f=/n/a/2006/07/09/sports/s134544D82.DTL&amp;feed=rss.news</feedburner:origLink>
+            </item>        
+    </channel>
+    </rss>
+    EOF
+      f = Syndication::RSS::Parser.new.parse(xml)
+      il = f.items
+      assert_not_nil(il)
+      assert(il.length == 1)
+      i = il.first
+      assert_not_nil(i.feedburner_origlink)
+      assert(i.feedburner_origlink == "http://www.sfgate.com/cgi-bin/article.cgi?f=/n/a/2006/07/09/sports/s134544D82.DTL&feed=rss.news")
+    end
+
+  end
+
+end

data/test/rsstest.rb

@@ -117,7 +117,6 @@ class Tests < Test::Unit::TestCase
           <link>http://www.example.com</link>
           <width>42</width>
           <height>23</height>
-          <description>The Example Logo</description>
         </image>
         <rating>(PICS-1.1 "http://www.icra.org/ratingsv02.html" l gen true r (cz 1 lz 1 nz 1 oz 1 vz 1) "http://www.rsac.org/ratingsv01.html" l gen true r (n 0 s 0 v 0 l 0) "http://www.classify.org/safesurf/" l gen true r (SS~~000 1))</rating>
         <textInput>
@@ -316,6 +315,108 @@ class Tests < Test::Unit::TestCase
     assert(i.itunes_duration == 232, "Duration computed incorrectly")
   end
 
+  # Test a well-formed RSS2 feed with every element possible and more than
+  # one item, with all kinds of stuff CDATA escaped.
+  def test_rss2_wf_full_cdata
+    xml = <<-EOF
+    <rss version="2">
+      <channel>
+        <title><![CDATA[Example Feed]]></title>
+        <link>http://www.example.com/</link>
+        <description><![CDATA[This is merely an example.]]></description>
+        <language>en-us</language>
+        <copyright>Copyright 2004 The Example Corporation.</copyright>
+        <managingEditor>editor@example.com</managingEditor>
+        <webMaster>webmaster@example.com</webMaster>
+        <pubDate>Sat, 07 Sep 2002 00:01:02 EDT</pubDate>
+        <lastBuildDate>Sat, 7 Sep 02 13:14:15 -0600</lastBuildDate>
+        <category>examples</category>
+        <category>boring</category>
+        <generator>vim of course</generator>
+        <docs>http://blogs.law.harvard.edu/tech/rss</docs>
+        <cloud domain="rpc.sys.com" port="80" path="/RPC2" registerProcedure="pingMe" protocol="soap"/>
+        <ttl>90</ttl>
+        <image>
+          <title><![CDATA[Example Inc]]></title>
+          <url>http://www.example.com/images/logo.jpg</url>
+          <link>http://www.example.com</link>
+          <width>42</width>
+          <height>23</height>
+        </image>
+        <rating>(PICS-1.1 "http://www.icra.org/ratingsv02.html" l gen true r (cz 1 lz 1 nz 1 oz 1 vz 1) "http://www.rsac.org/ratingsv01.html" l gen true r (n 0 s 0 v 0 l 0) "http://www.classify.org/safesurf/" l gen true r (SS~~000 1))</rating>
+        <textInput>
+          <title>Submit</title>
+          <description>Enter keywords</description>
+          <name>SearchKeywords</name>
+          <link>http://www.example.com/cgi-bin/search.pl</link>
+        </textInput>
+        <skipHours>
+          <hour>0</hour>
+          <hour>23</hour>
+        </skipHours>
+        <skipDays>
+          <day>Monday</day>
+          <day>Sunday</day>
+        </skipDays>
+        <item>
+          <title>Our stock price shot up</title>
+          <link>http://www.example.com/news/2.html</link>
+          <description>We were hyped in the press!</description>
+        </item>
+        <item>
+          <title><![CDATA[Unencoded < > and & are allowed.]]></title>
+          <link><![CDATA[http://www.example.com/news/1.html]]></link>
+          <description><![CDATA[If this was any less interesting, it would be amazing.]]></description>
+          <author><![CDATA[fred@example.com]]></author>
+          <pubDate>Sat, 07 Sep 2002 00:01:02 EDT</pubDate>
+          <category>dull</category>
+          <category>amazingly</category>
+          <comments>http://www.example.com/news/comments/1.html</comments>
+          <enclosure url="http://www.example.com/mp3/advertisement.mp3" length="123987" type="audio/mpeg" />
+          <guid>4asd98dgf9a74@example.com</guid>
+          <source url="http://www.example.com/news.xml">Example News</source>
+        </item>
+      </channel>
+    </rss>
+    EOF
+    f = Syndication::RSS::Parser.new.parse(xml)
+    baseline_rss_assertions(f)
+    for elem in %w(title link description language copyright managingeditor webmaster pubdate lastbuilddate category generator docs cloud ttl textinput rating skiphours skipdays)
+      assert_not_nil(f.channel.send(elem), "feed.channel.#{elem} is nil, it shouldn't be")
+      assert(f.channel.send(elem).to_s.length > 0)
+    end
+    # Check CDATA is decoded properly
+    assert(f.channel.title == 'Example Feed')
+    assert(f.channel.description == 'This is merely an example.')
+    items = f.items
+    assert(items.length == 2)
+    i = items.last
+    for elem in %w(title link description author pubdate category comments enclosure guid source)
+      assert_not_nil(i.send(elem), "feed.channel.item[1].#{elem} is nil, it shouldn't be")
+    end
+    cats = i.category
+    assert(i.title == 'Unencoded < > and & are allowed.')
+    assert(i.link == 'http://www.example.com/news/1.html')
+    assert(cats.length == 2)
+    assert(cats.first == 'dull')
+    assert(cats.last == 'amazingly')
+    assert(f.channel.skiphours.length == 2)
+    assert(f.channel.skiphours.first == 0)
+    assert(f.channel.skiphours.last == 23)
+    assert(f.channel.pubdate.kind_of?(DateTime))
+    assert(f.channel.lastbuilddate.kind_of?(DateTime))
+    assert(f.channel.pubdate.mday == 7)
+    assert(f.channel.pubdate.month == 9)
+    assert(f.channel.lastbuilddate.mday == 7)
+    assert(f.channel.lastbuilddate.month == 9)
+    c = f.channel
+    assert_not_nil(c)
+    assert_kind_of(Syndication::RSS::Channel, c)
+    assert_not_nil(c.title)
+    assert_not_nil(c.link)
+    assert_not_nil(c.description)
+  end
+
 end
 
 end

metadata

@@ -3,8 +3,8 @@ rubygems_version: 0.8.11
 specification_version: 1
 name: syndication
 version: !ruby/object:Gem::Version 
-  version: "0.6"
-date: 2006-06-18 00:00:00 -05:00
+  version: 0.6.1
+date: 2006-07-17 00:00:00 -05:00
 summary: A web syndication parser for Atom and RSS with a uniform API
 require_paths: 
 - lib
@@ -38,6 +38,8 @@ files:
 - lib/syndication/rss.rb
 - lib/syndication/syndication.rb
 - lib/syndication/atom.rb
+- lib/syndication/feedburner.rb
+- test/feedburntest.rb
 - test/tagsouptest.rb
 - test/google.rb
 - test/rsstest.rb
@@ -55,6 +57,7 @@ test_files:
 - test/rsstest.rb
 - test/google.rb
 - test/tagsouptest.rb
+- test/feedburntest.rb
 rdoc_options: []
 
 extra_rdoc_files: