rrsimple-rss 1.3.5 → 1.3.6

checksums.yaml

@@ -0,0 +1,15 @@
+---
+!binary "U0hBMQ==":
+  metadata.gz: !binary |-
+    YTg0ZWNjZGZjZWU5NmVkODVmOWU1ZGNkM2NlYWE4ZDZkODZlZjc2Mg==
+  data.tar.gz: !binary |-
+    Y2I4NWY4MjlhMDM3MTE1Y2FmNjBlZTEzNTM2YTFjZmI0ZjkwMjlkOA==
+SHA512:
+  metadata.gz: !binary |-
+    ZjU0YTA3YWM1M2Y0ZjM4YWIwZmUyMTliNTUyYzg3YmI5YjRiMWY5MmQ0M2Q1
+    Mjc1NDkyZDQ0NmY4OTIwM2NjNTRhMzA4N2VlZjRiYTcwYzIxN2Y2YTQ3M2Q3
+    MTdkOWNlYmUyYmFjNjI0MDg3MjdhOGU4ZjM3MjhiMDY4MGFmMTg=
+  data.tar.gz: !binary |-
+    ZjRiMTA1YzhlZTg1NzczM2JlMjVkNjIwNjAwMmNmZTg0N2Y1MzY1MTJhODY2
+    NDE1MzIyODI2MGNjYWRjMzNjZDcxZjA1ZmEyOTEwODQwOTM0N2I4YzAyODhh
+    MTUyOGFjN2I5YTNlYWYyYjFkMDIxNzlhYWI4ODBkNTQ2ZDY3ZTY=

data/lib/rrsimple-rss.rb

@@ -2,17 +2,17 @@ require 'cgi'
 require 'time'
 
 class RRSimpleRSS
-  VERSION = "1.3.5"
-  
+  VERSION = "1.3.6"
+
   attr_reader :items, :source
   alias :entries :items
 
   @@feed_tags = [
     :id,
     :title, :subtitle, :link,
-    :description, 
+    :description,
     :author, :webMaster, :managingEditor, :contributor,
-    :pubDate, :lastBuildDate, :updated, :'dc:date',
+    :pubDate, :lastBuildDate, :updated, :'dc:date', :'dc:subject',
     :generator, :language, :docs, :cloud,
     :ttl, :skipHours, :skipDays,
     :image, :logo, :icon, :rating,
@@ -26,7 +26,7 @@ class RRSimpleRSS
     :title, :link, :'link+alternate', :'link+self', :'link+edit', :'link+replies',
     :author, :contributor,
     :description, :summary, :content, :'content:encoded', :comments,
-    :pubDate, :published, :updated, :expirationDate, :modified, :'dc:date',
+    :pubDate, :published, :updated, :expirationDate, :modified, :'dc:date', :'dc:subject',
     :category, :categories, :guid,
     :'trackback:ping', :'trackback:about',
     :'dc:creator', :'dc:title', :'dc:subject', :'dc:rights', :'dc:publisher',
@@ -41,7 +41,7 @@ class RRSimpleRSS
     @source = source.respond_to?(:read) ? source.read : source.to_s
     @items = Array.new
     @options = Hash.new.update(options)
-    
+
     parse
   end
 
@@ -73,10 +73,10 @@ class RRSimpleRSS
 
   def parse
     raise RRSimpleRSSError, "Poorly formatted feed" unless @source =~ %r{<(channel|feed).*?>.*?</(channel|feed)>}mi
-    
+
     # Feed's title and link
     feed_content = $1 if @source =~ %r{(.*?)<(rss:|atom:)?(item|entry).*?>.*?</(rss:|atom:)?(item|entry)>}mi
-    
+
     @@feed_tags.each do |tag|
       if feed_content && feed_content =~ %r{<(rss:|atom:)?#{tag}(.*?)>(.*?)</(rss:|atom:)?#{tag}>}mi
         nil
@@ -87,14 +87,14 @@ class RRSimpleRSS
       elsif @source =~ %r{<(rss:|atom:)?#{tag}(.*?)\/\s*>}mi
         nil
       end
-      
+
       if $2 || $3
         tag_cleaned = clean_tag(tag)
         instance_variable_set("@#{ tag_cleaned }", clean_content(tag, $2, $3))
         self.class.class_eval("attr_reader :#{ tag_cleaned }")
       end
     end
-    
+
     # RSS items' title, link, and description
     @source.scan( %r{<(rss:|atom:)?(item|entry)([\s][^>]*)?>(.*?)</(rss:|atom:)?(item|entry)>}mi ) do |match|
       item = Hash.new
@@ -103,13 +103,13 @@ class RRSimpleRSS
           tag_data = tag.to_s.split("+")
           tag = tag_data[0]
           rel = tag_data[1]
-          
+
           if match[3] =~ %r{<(rss:|atom:)?#{tag}(.*?)rel=['"]#{rel}['"](.*?)>(.*?)</(rss:|atom:)?#{tag}>}mi
             nil
           elsif match[3] =~ %r{<(rss:|atom:)?#{tag}(.*?)rel=['"]#{rel}['"](.*?)/\s*>}mi
             nil
           end
-          
+
           item[clean_tag("#{tag}+#{rel}")] = clean_content(tag, $3, $4) if $3 || $4
         elsif tag.to_s.include?("#")
           tag_data = tag.to_s.split("#")
@@ -120,7 +120,7 @@ class RRSimpleRSS
           elsif match[3] =~ %r{<(rss:|atom:)?#{tag}(.*?)#{attrib}=['"](.*?)['"](.*?)/\s*>}mi
             nil
           end
-          
+
           item[clean_tag("#{tag}_#{attrib}")] = clean_content(tag, attrib, $3) if $3
         else
           if match[3] =~ %r{<(rss:|atom:)?#{tag}(.*?)>(.*?)</(rss:|atom:)?#{tag}>}mi
@@ -128,13 +128,15 @@ class RRSimpleRSS
           elsif match[3] =~ %r{<(rss:|atom:)?#{tag}(.*?)/\s*>}mi
             nil
           end
-          
-          if tag == :category || tag == :keywords
+
+          if tag == :category
             item[tag] = match[3].scan(%r{<(rss:|atom:)?#{tag}(.*?)>(.*?)</(rss:|atom:)?#{tag}>}mi).map {|x| category(x[2])}
+          elsif tag == :keywords
+            item[tag] = match[3].scan(%r{<(rss:|atom:)?dc:subject rdf:type="keywords"(.*?)>(.*?)<\/(rss:|atom:)?dc:subject>}mi).map {|x| x[2].gsub(";", ",") }
           else
             item[clean_tag(tag)] = clean_content(tag, $2, $3) if $2 || $3
           end
-          
+
         end
       end
 
@@ -157,7 +159,10 @@ class RRSimpleRSS
       when :author, :contributor, :skipHours, :skipDays
         unescape(content.gsub(/<.*?>/,''))
       when :"full-text"
-        CGI.unescapeHTML unescape(content.gsub(/<.*?>/,'')).force_encoding("UTF-8")        
+        CGI.unescapeHTML unescape(content.gsub(/<.*?>/,'')).force_encoding("UTF-8")
+      when :link
+        "#{attrs} " =~ /href=['"]?([^'"]*)['" ]/mi
+        $1
       else
         content.empty? && "#{attrs} " =~ /href=['"]?([^'"]*)['" ]/mi ? $1.strip : unescape(content)
     end
@@ -166,7 +171,7 @@ class RRSimpleRSS
   def clean_tag(tag)
     tag != :'full-text' ? tag.to_s.gsub(':','_').intern : tag.to_s.gsub('-','_').intern
   end
-  
+
   def unescape(content)
     if content.respond_to?(:force_encoding) && content.force_encoding("binary") =~ /([^-_.!~*'()a-zA-Z\d;\/?:@&=+$,\[\]]%)/n then
       CGI.unescape(content).gsub(/(<!\[CDATA\[|\]\]>)/,'').strip

data/rrsimple-rss.gemspec

@@ -1,6 +1,6 @@
 Gem::Specification.new do |s|
   s.name = "rrsimple-rss"
-  s.version = "1.3.5"
+  s.version = "1.3.6"
   s.date = "2010-07-06"
   s.summary = "A simple, flexible, extensible, and liberal RSS and Atom reader for Ruby. It is designed to be backwards compatible with the standard RSS parser, but will never do RSS generation."
   s.email = "rodrigo@rrmartins.com"

data/test/base/base_test.rb

@@ -7,45 +7,48 @@ require 'pry'
 class BaseTest < Test::Unit::TestCase
   def setup
     @rss_prnews = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/prnews.xml')
-    @rss_bbc = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/bbc.xml')
-    @rss_baboo = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/baboo.xml')
-    @rss09 = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/rss09.rdf')
-    @rss20 = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/rss20.xml')
-    @media_rss = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/media_rss.xml')
-    @atom = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/atom.xml')
+    @rss_bbc    = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/bbc.xml')
+    @rss_bbc2   = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/bbc2.xml')
+    @rss_baboo  = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/baboo.xml')
+    @rss09      = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/rss09.rdf')
+    @rss20      = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/rss20.xml')
+    @media_rss  = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/media_rss.xml')
+    @atom       = RRSimpleRSS.parse open(File.dirname(__FILE__) + '/../data/atom.xml')
   end
-  
+
   def test_channel
-    assert_equal @rss_bbc, @rss_bbc.channel
-    assert_equal @rss09, @rss09.channel
-    assert_equal @rss20, @rss20.channel
-    assert_equal @atom, @atom.feed
-    assert_equal @rss_baboo, @rss_baboo.feed
+    assert_equal @rss_bbc2,   @rss_bbc2.channel
+    assert_equal @rss_bbc,    @rss_bbc.channel
+    assert_equal @rss09,      @rss09.channel
+    assert_equal @rss20,      @rss20.channel
+    assert_equal @atom,       @atom.feed
+    assert_equal @rss_baboo,  @rss_baboo.feed
     assert_equal @rss_prnews, @rss_prnews.feed
   end
-  
+
   def test_items
     assert_kind_of Array, @rss_baboo.items
     assert_kind_of Array, @rss_bbc.items
+    assert_kind_of Array, @rss_bbc2.items
     assert_kind_of Array, @rss09.items
     assert_kind_of Array, @rss20.items
     assert_kind_of Array, @atom.entries
     assert_kind_of Array, @rss_prnews.items
   end
-  
+
   def test_rss_prnews
-    assert_equal 20, @rss_prnews.items.size
+    assert_equal 20,                     @rss_prnews.items.size
     assert_equal "PR Newswire Releases", @rss_prnews.title.force_encoding("UTF-8")
     assert_equal "http://www.prnewswire.com/news-releases/sva-promove-exposicao-de-cartazes-que-retrata-a-evolucao-do-design-grafico-nos-ultimos-70-anos-540199401.html", @rss_prnews.channel.link
     assert_equal "http://www.prnewswire.com/news-releases/sva-promove-exposicao-de-cartazes-que-retrata-a-evolucao-do-design-grafico-nos-ultimos-70-anos-540199401.html", @rss_prnews.items.first.link
     assert_equal "http://www.prnewswire.com/news-releases/sva-promove-exposicao-de-cartazes-que-retrata-a-evolucao-do-design-grafico-nos-ultimos-70-anos-540199401.html", @rss_prnews.items.first[:link]
     assert_equal Time.parse("2015-11-04 15:30:00 -0200"), @rss_prnews.items.first.pubDate
     assert_equal Time.parse("2015-11-09 11:55:45 -0200"), @rss_prnews.channel.lastBuildDate
-    assert_nil @rss_prnews.items.first.full_text
+    assert_nil     @rss_prnews.items.first.full_text
     assert_not_nil @rss_prnews.items.first.description
     assert_kind_of Array, @rss_prnews.items.first.keywords
     assert_kind_of Array, @rss_prnews.items.first.category
-    assert_equal "Art", @rss_prnews.items.first.category.first
+    assert_equal "Art",   @rss_prnews.items.first.category.first
   end
 
   def test_rss_baboo
@@ -56,8 +59,8 @@ class BaseTest < Test::Unit::TestCase
     assert_equal "http://www.baboo.com.br/internet/lista-do-procon-sp-com-lojas-online-que-devem-ser-evitadas-e-atualizada/", @rss_baboo.items.first[:link]
     assert_equal Time.parse("2015-11-05 14:11:24 -0200"), @rss_baboo.items.first.pubDate
     assert_equal Time.parse("2015-11-05 14:11:24 -0200"), @rss_baboo.channel.lastBuildDate
-    assert_nil @rss_baboo.items.first.full_text
-    assert_not_nil @rss_baboo.items.first.description
+    assert_nil            @rss_baboo.items.first.full_text
+    assert_not_nil        @rss_baboo.items.first.description
     assert_kind_of Array, @rss_baboo.items.first.category
   end
 
@@ -73,6 +76,19 @@ class BaseTest < Test::Unit::TestCase
     assert_kind_of Array, @rss_bbc.items.first.category
   end
 
+  def test_rss_bbc2
+    assert_equal 3, @rss_bbc2.items.size
+    assert_equal "BBCBrasil.com | Notícias", @rss_bbc2.title.force_encoding("UTF-8")
+    assert_equal "http://www.bbc.com/portuguese/full_all.xml", @rss_bbc2.channel.link
+    assert_equal "http://www.bbc.com/portuguese/noticias/2015/11/151109_catacumbas_paris_brasileiro_cc", @rss_bbc2.items.first.link
+    assert_equal "http://www.bbc.com/portuguese/noticias/2015/11/151109_catacumbas_paris_brasileiro_cc", @rss_bbc2.items.first[:link]
+    assert_equal Time.parse("2015-11-09 16:40:33 -0200"), @rss_bbc2.items.first.updated
+    assert_equal Time.parse("2015-11-09 13:29:35 -0200"), @rss_bbc2.channel.updated
+    assert_not_nil @rss_bbc2.items.first.content
+    assert_nil @rss_bbc2.items.first.category.first
+    assert_kind_of Array, @rss_bbc2.items.first.keywords
+  end
+
   def test_rss09
     assert_equal 10, @rss09.items.size
     assert_equal "Slashdot", @rss09.title
@@ -106,7 +122,7 @@ class BaseTest < Test::Unit::TestCase
     assert_nil @media_rss.items.first.full_text
     assert_kind_of Array, @media_rss.items.first.category
   end
-  
+
   def test_rss20
     assert_equal 10, @rss20.items.size
     assert_equal "Technoblog", @rss20.title
@@ -117,7 +133,7 @@ class BaseTest < Test::Unit::TestCase
     assert_nil @rss20.items.first.full_text
     assert_kind_of Array, @rss20.items.first.category
   end
-  
+
   def test_atom
     assert_equal 1, @atom.entries.size
     assert_equal "dive into mark", @atom.title
@@ -127,7 +143,7 @@ class BaseTest < Test::Unit::TestCase
     assert_nil @atom.entries.first.full_text
     assert_kind_of Array, @atom.entries.first.category
   end
-  
+
   def test_bad_feed
     assert_raise(RRSimpleRSSError) { RRSimpleRSS.parse(open(File.dirname(__FILE__) + '/../data/not-rss.xml')) }
   end

metadata

@@ -1,8 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: rrsimple-rss
 version: !ruby/object:Gem::Version
-  version: 1.3.5
-  prerelease: 
+  version: 1.3.6
 platform: ruby
 authors:
 - Lucas Carlson
@@ -20,42 +19,41 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
-- install.rb
-- lib/rrsimple-rss.rb
 - LICENSE
-- Rakefile
 - README.markdown
+- Rakefile
+- install.rb
+- lib/rrsimple-rss.rb
 - rrsimple-rss.gemspec
 - test/base/base_test.rb
-- test/data/bbc.xml
 - test/data/atom.xml
+- test/data/bbc.xml
 - test/data/not-rss.xml
 - test/data/rss09.rdf
 - test/data/rss20.xml
 - test/test_helper.rb
 homepage: http://github.com/rrmartins/simple-rss
 licenses: []
+metadata: {}
 post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
-  none: false
   requirements:
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
-  none: false
   requirements:
   - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
 rubyforge_project: rrsimple-rss
-rubygems_version: 1.8.23.2
+rubygems_version: 2.4.8
 signing_key: 
-specification_version: 3
+specification_version: 4
 summary: A simple, flexible, extensible, and liberal RSS and Atom reader for Ruby.
   It is designed to be backwards compatible with the standard RSS parser, but will
   never do RSS generation.