feedtools 0.2.15 → 0.2.16

data/CHANGELOG

@@ -1,3 +1,10 @@
+== FeedTools 0.2.16
+ * atom feed generation now defaults to atom 1.0
+ * removed erroneous published element at atom feed-level
+ * fixed some incompatibilities with ruby 1.8.3
+ * better support for universal subscription mechanism within feeds
+ * now handles both atom 0.3 and 1.0 namespaces
+ * fixed timestamping of feed items
 == FeedTools 0.2.15
  * changed so that tidied elements no longer handle entities differently
  * tidy now assumes utf-8 encoding instead of ascii

data/lib/feed_tools.rb

@@ -32,13 +32,14 @@ FEED_TOOLS_ENV = ENV['FEED_TOOLS_ENV'] ||
                  ENV['RAILS_ENV'] ||
                  'production' # :nodoc:
 
-FEED_TOOLS_VERSION = "0.2.15"
+FEED_TOOLS_VERSION = "0.2.16"
 
 FEED_TOOLS_NAMESPACES = {
   "admin" => "http://webns.net/mvcb/",
   "ag" => "http://purl.org/rss/1.0/modules/aggregation/",
   "annotate" => "http://purl.org/rss/1.0/modules/annotate/",
   "atom" => "http://www.w3.org/2005/Atom",
+  "atom03" => "http://purl.org/atom/ns#",
   "audio" => "http://media.tangent.org/rss/1.0/",
   "blogChannel" => "http://backend.userland.com/blogChannelModule",
   "cc" => "http://web.resource.org/cc/",
@@ -81,6 +82,9 @@ FEED_TOOLS_NAMESPACES = {
   "xml" => "http://www.w3.org/XML/1998/namespace"
 }
 
+FEED_TOOLS_NAMESPACES_WITH_OLD_ATOM = FEED_TOOLS_NAMESPACES.clone
+FEED_TOOLS_NAMESPACES_WITH_OLD_ATOM['atom'] = "http://purl.org/atom/ns#"
+
 $:.unshift(File.dirname(__FILE__))
 $:.unshift(File.dirname(__FILE__) + "/feed_tools/vendor")
 
@@ -508,6 +512,7 @@ module FeedTools
       end
       tidy_html.gsub!(/&/, "&")
       tidy_html.gsub!(/&/, "&")
+      tidy_html.gsub!(/\320\262\320\202\342\204\242/, "\342\200\231")
     else
       tidy_html = html
     end

data/lib/feed_tools/feed.rb

@@ -185,7 +185,12 @@ module FeedTools
       end
     
       # Find out what method we're going to be using to obtain this feed.
-      uri = URI.parse(self.url)
+      begin
+        uri = URI.parse(self.url)
+      rescue URI::InvalidURIError
+        raise FeedAccessError,
+          "Cannot retrieve feed using invalid URL: " + self.url.to_s
+      end
       retrieval_method = "http"
       case uri.scheme
       when "http"
@@ -352,8 +357,8 @@ module FeedTools
           end
           unless @http_response.kind_of? Net::HTTPRedirection
             @http_headers = {}
-            self.http_response.each_header do |header|
-              self.http_headers[header.first.downcase] = header.last
+            self.http_response.each_header do |key, value|
+              self.http_headers[key.downcase] = value
             end
             self.last_retrieved = Time.now
             self.feed_data = self.http_response.body
@@ -392,8 +397,8 @@ module FeedTools
           end
           if @http_response != nil
             @http_headers = {}
-            self.http_response.each_header do |header|
-              self.http_headers[header.first] = header.last
+            self.http_response.each_header do |key, value|
+              self.http_headers[key.downcase] = value
             end
             if self.http_response.code.to_i == 304
               self.last_retrieved = Time.now
@@ -680,6 +685,10 @@ module FeedTools
       if @id.nil?
         unless channel_node.nil?
           @id = XPath.first(channel_node, "id/text()").to_s
+          if @id == ""
+            @id = XPath.first(channel_node, "atom:id/text()",
+              FEED_TOOLS_NAMESPACES).to_s
+          end
           if @id == ""
             @id = XPath.first(channel_node, "guid/text()").to_s
           end
@@ -688,6 +697,10 @@ module FeedTools
           if @id == "" || @id.nil?
             @id = XPath.first(root_node, "id/text()").to_s
           end
+          if @id == ""
+            @id = XPath.first(root_node, "atom:id/text()",
+              FEED_TOOLS_NAMESPACES).to_s
+          end
           if @id == ""
             @id = XPath.first(root_node, "guid/text()").to_s
           end
@@ -704,9 +717,66 @@ module FeedTools
   
     # Returns the feed url.
     def url
-      if @url.nil? && self.feed_data != nil
+      original_url = @url
+      override_url = lambda do
+        begin
+          if @url == nil && self.feed_data != nil
+            true
+          elsif @url != nil &&
+              !(["http", "https"].include?(URI.parse(@url).scheme))
+            if self.feed_data != nil
+              true
+            else
+              false
+            end
+          else
+            false
+          end
+        rescue
+          true
+        end
+      end
+      if override_url.call
         @url = XPath.first(channel_node, "link[@rel='self']/@href").to_s
         @url = nil if @url == ""
+        if override_url.call
+          @url = XPath.first(channel_node, "atom:link[@rel='self']/@href").to_s
+          @url = nil if @url == ""
+        end
+        if override_url.call
+          @url = XPath.first(channel_node, "atom:link[@rel='self']/@href",
+            FEED_TOOLS_NAMESPACES).to_s
+          @url = nil if @url == ""
+        end
+        if override_url.call
+          @url = XPath.first(channel_node, "atom:link[@rel='self']/@href",
+            FEED_TOOLS_NAMESPACES_WITH_OLD_ATOM).to_s
+          @url = nil if @url == ""
+        end
+        if override_url.call
+          @url = XPath.first(channel_node, "admin:feed/@rdf:resource").to_s
+          @url = nil if @url == ""
+        end
+        if override_url.call
+          @url = XPath.first(channel_node, "admin:feed/@rdf:resource",
+            FEED_TOOLS_NAMESPACES).to_s
+          @url = nil if @url == ""
+        end
+        if override_url.call
+          @url = XPath.first(channel_node, "admin:feed/@resource").to_s
+          @url = nil if @url == ""
+        end
+        if override_url.call
+          @url = XPath.first(channel_node, "feed/@rdf:resource").to_s
+          @url = nil if @url == ""
+        end
+        if override_url.call
+          @url = XPath.first(channel_node, "feed/@resource").to_s
+          @url = nil if @url == ""
+        end
+        if @url == nil
+          @url = original_url
+        end
       end
       return @url
     end
@@ -1213,7 +1283,7 @@ module FeedTools
       return @itunes_author
     end
 
-    # Returns the feed item time
+    # Returns the feed time
     def time
       if @time.nil?
         unless channel_node.nil?
@@ -1233,11 +1303,8 @@ module FeedTools
         end
         begin
           if time_string != nil && time_string != ""
-            @time = Time.parse(time_string) rescue self.succ_time
-          elsif time_string == nil
-            @time = self.succ_time
-          end
-          if @time == nil
+            @time = Time.parse(time_string)
+          else
             @time = Time.now
           end
         rescue
@@ -1252,29 +1319,6 @@ module FeedTools
       @time = new_time
     end
   
-    # Returns 1 second after the previous item's time.
-    def succ_time #:nodoc:
-      begin
-        if feed.nil?
-          return nil
-        end
-        feed.items
-        unsorted_items = feed.instance_variable_get("@items")
-        item_index = unsorted_items.index(self)
-        if item_index.nil?
-          return nil
-        end
-        if item_index <= 0
-          return Time.now
-        end
-        previous_item = unsorted_items[item_index - 1]
-        return previous_item.time.succ
-      rescue
-        return nil
-      end
-    end
-    private :succ_time
-
     # Returns the feed item updated time
     def updated
       if @updated.nil?
@@ -1729,7 +1773,7 @@ module FeedTools
         # create the individual feed items
         @items = []
         if raw_items != nil
-          for item_node in raw_items
+          for item_node in raw_items.reverse
             new_item = FeedItem.new
             new_item.feed_data = item_node.to_s
             new_item.feed_data_type = self.feed_data_type
@@ -1816,7 +1860,7 @@ module FeedTools
       if feed_type == "rss" && (version == nil || version == 0.0)
         version = 1.0
       elsif feed_type == "atom" && (version == nil || version == 0.0)
-        version = 0.3
+        version = 1.0
       end
       if feed_type == "rss" && (version == 0.9 || version == 1.0 ||
           version == 1.1)
@@ -2030,9 +2074,6 @@ module FeedTools
           else
             xml_builder.updated(Time.now.iso8601)
           end
-          unless self.published.nil?
-            xml_builder.published(self.published.iso8601)            
-          end
           xml_builder.generator("FeedTools - " +
             "http://www.sporkmonger.com/projects/feedtools")
           if self.id != nil

data/lib/feed_tools/feed_item.rb

@@ -1266,9 +1266,17 @@ module FeedTools
             time_string = XPath.first(root_node, "time/text()").to_s
           end
         end
-        if time_string != nil && time_string != ""
-          @time = Time.parse(time_string) rescue Time.now
-        elsif time_string == nil
+        begin
+          time_string = "" if time_string.nil?
+          if time_string != ""
+            @time = Time.parse(time_string)
+          else
+            @time = succ_time
+          end
+        rescue
+          @time = succ_time
+        end
+        if @time.nil?
           @time = Time.now
         end
       end
@@ -1279,6 +1287,31 @@ module FeedTools
     def time=(new_time)
       @time = new_time
     end
+    
+    # Returns 1 second after the previous item's time.
+    def succ_time #:nodoc:
+      begin
+        if feed.nil?
+          return nil
+        end
+        if feed.instance_variable_get("@items").nil?
+          feed.items
+        end
+        unsorted_items = feed.instance_variable_get("@items")
+        item_index = unsorted_items.index(self)
+        if item_index.nil?
+          return nil
+        end
+        if item_index <= 0
+          return Time.now
+        end
+        previous_item = unsorted_items[item_index - 1]
+        return previous_item.time.succ
+      rescue
+        return nil
+      end
+    end
+    private :succ_time
 
     # Returns the feed item updated time
     def updated
@@ -1462,7 +1495,7 @@ module FeedTools
       if feed_type == "rss" && (version == nil || version == 0.0)
         version = 1.0
       elsif feed_type == "atom" && (version == nil || version == 0.0)
-        version = 0.3
+        version = 1.0
       end
       if feed_type == "rss" && (version == 0.9 || version == 1.0 || version == 1.1)
         # RDF-based rss format
@@ -1663,4 +1696,4 @@ module FeedTools
         "LINK:#{self.link}>"
     end
   end
-end
+end

data/rakefile

@@ -7,7 +7,7 @@ require 'rake/gempackagetask'
 require 'rake/contrib/rubyforgepublisher'
 
 PKG_NAME      = 'feedtools'
-PKG_VERSION   = '0.2.15'
+PKG_VERSION   = '0.2.16'
 PKG_FILE_NAME = "#{PKG_NAME}-#{PKG_VERSION}"
 
 RELEASE_NAME  = "REL #{PKG_VERSION}"

data/test/rss_test.rb

@@ -581,9 +581,58 @@ class RssTest < Test::Unit::TestCase
             </content:encoded>
           </item>
         </channel>
-      </feed>
+      </rss>
     FEED
     assert_equal("Test Feed Title", feed.items.first.title)
     assert_equal("Test Feed Content", feed.items.first.content)
   end
+  
+  def test_item_order
+    feed = FeedTools::Feed.new
+    feed.feed_data = <<-FEED
+      <rss>
+        <channel>
+          <item>
+            <title>Item 1</title>
+          </item>
+          <item>
+            <title>Item 2</title>
+          </item>
+          <item>
+            <title>Item 3</title>
+          </item>
+        </channel>
+      </rss>
+    FEED
+    assert_equal("Item 1", feed.items[0].title)
+    assert_equal("Item 2", feed.items[1].title)
+    assert_equal("Item 3", feed.items[2].title)
+  end
+  
+  def test_usm
+    feed = FeedTools::Feed.new
+    feed.feed_data = <<-FEED
+      <rss xmlns:atom="http://www.w3.org/2005/Atom">
+        <channel>
+          <atom:link href="http://nowhere.com/feed.xml" rel="self" />
+          <title>Feed Title</title>
+        </channel>
+      </rss>
+    FEED
+    assert_equal("http://nowhere.com/feed.xml", feed.url)
+    assert_equal("Feed Title", feed.title)
+    
+    feed = FeedTools::Feed.new
+    feed.feed_data = <<-FEED
+      <rss xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
+           xmlns:admin="http://webns.net/mvcb/">
+        <channel>
+          <admin:feed rdf:resource="http://nowhere.com/feed.xml" />
+          <title>Feed Title</title>
+        </channel>
+      </rss>
+    FEED
+    assert_equal("http://nowhere.com/feed.xml", feed.url)
+    assert_equal("Feed Title", feed.title)
+  end
 end

metadata

@@ -3,8 +3,8 @@ rubygems_version: 0.8.11
 specification_version: 1
 name: feedtools
 version: !ruby/object:Gem::Version 
-  version: 0.2.15
-date: 2005-10-01 00:00:00 -04:00
+  version: 0.2.16
+date: 2005-10-18 00:00:00 -04:00
 summary: "Parsing, generation, and caching system for xml news feeds."
 require_paths: 
   - lib