simple-rss 1.3.2 → 2.0.0

data/test/base/empty_tag_test.rb

@@ -0,0 +1,56 @@
+require_relative "../test_helper"
+require "timeout"
+
+class EmptyTagTest < Test::Unit::TestCase
+  def setup
+    @rss20 = SimpleRSS.parse(open(File.dirname(__FILE__) + "/../data/rss20.xml"))
+  end
+
+  def test_empty_item_tag_does_not_hang
+    # Reproduces issue #16: 100% cpu and hanging process on blank item tag
+    # Adding an empty tag should not cause regex catastrophic backtracking
+    original_tags = SimpleRSS.item_tags.dup
+
+    begin
+      SimpleRSS.item_tags << :""
+
+      # This should complete quickly, not hang
+      Timeout.timeout(5) do
+        rss = SimpleRSS.parse(open(File.dirname(__FILE__) + "/../data/rss20.xml"))
+        assert_not_nil rss.items
+      end
+    ensure
+      SimpleRSS.item_tags.replace(original_tags)
+    end
+  end
+
+  def test_blank_item_tag_does_not_hang
+    original_tags = SimpleRSS.item_tags.dup
+
+    begin
+      SimpleRSS.item_tags << :"   "
+
+      Timeout.timeout(5) do
+        rss = SimpleRSS.parse(open(File.dirname(__FILE__) + "/../data/rss20.xml"))
+        assert_not_nil rss.items
+      end
+    ensure
+      SimpleRSS.item_tags.replace(original_tags)
+    end
+  end
+
+  def test_empty_feed_tag_does_not_hang
+    original_tags = SimpleRSS.feed_tags.dup
+
+    begin
+      SimpleRSS.feed_tags << :""
+
+      Timeout.timeout(5) do
+        rss = SimpleRSS.parse(open(File.dirname(__FILE__) + "/../data/rss20.xml"))
+        assert_not_nil rss.channel
+      end
+    ensure
+      SimpleRSS.feed_tags.replace(original_tags)
+    end
+  end
+end

data/test/base/encoding_test.rb

@@ -0,0 +1,87 @@
+require_relative "../test_helper"
+
+class EncodingTest < Test::Unit::TestCase
+  def test_strings_are_utf8_encoded
+    rss = SimpleRSS.parse(open(File.dirname(__FILE__) + "/../data/rss20.xml"))
+
+    assert_equal Encoding::UTF_8, rss.title.encoding
+    assert_equal Encoding::UTF_8, rss.link.encoding
+    assert_equal Encoding::UTF_8, rss.description.encoding
+  end
+
+  def test_strings_without_percent_are_utf8_encoded
+    # Issue #28: strings without '%' were returning ASCII-8BIT
+    rss = SimpleRSS.parse(open(File.dirname(__FILE__) + "/../data/rss20.xml"))
+
+    rss.items.each do |item|
+      assert_equal Encoding::UTF_8, item.title.encoding, "Item title should be UTF-8"
+      assert_equal Encoding::UTF_8, item.link.encoding, "Item link should be UTF-8"
+    end
+  end
+
+  def test_utf8_content_preserved
+    rss = SimpleRSS.parse(open(File.dirname(__FILE__) + "/../data/rss20_utf8.xml"))
+
+    assert_equal Encoding::UTF_8, rss.title.encoding
+    # Verify UTF-8 characters are preserved
+    assert(rss.items.any? { |item| item.title && item.title.encoding == Encoding::UTF_8 })
+  end
+
+  def test_ascii_8bit_source_normalized_to_utf8
+    # Issue #28: when source is ASCII-8BIT, output should still be UTF-8
+    xml = <<~XML.b # .b forces ASCII-8BIT encoding
+      <?xml version="1.0"?>
+      <rss version="2.0">
+        <channel>
+          <title>Test Feed</title>
+          <link>http://example.com</link>
+          <description>A test feed</description>
+          <item>
+            <title>Test Item</title>
+            <link>http://example.com/item</link>
+          </item>
+        </channel>
+      </rss>
+    XML
+
+    assert_equal Encoding::ASCII_8BIT, xml.encoding, "Source should be ASCII-8BIT"
+
+    rss = SimpleRSS.parse(xml)
+
+    assert_equal Encoding::UTF_8, rss.title.encoding, "Title should be UTF-8"
+    assert_equal Encoding::UTF_8, rss.link.encoding, "Link should be UTF-8"
+    assert_equal Encoding::UTF_8, rss.items.first.title.encoding, "Item title should be UTF-8"
+  end
+
+  def test_consistent_encoding_with_and_without_percent
+    # Issue #28: CGI.unescape returns UTF-8, but without '%' we got ASCII-8BIT
+    xml_with_percent = <<~XML.b
+      <?xml version="1.0"?>
+      <rss version="2.0">
+        <channel>
+          <title>Test%20Feed</title>
+          <link>http://example.com</link>
+          <description>Test</description>
+        </channel>
+      </rss>
+    XML
+
+    xml_without_percent = <<~XML.b
+      <?xml version="1.0"?>
+      <rss version="2.0">
+        <channel>
+          <title>Test Feed</title>
+          <link>http://example.com</link>
+          <description>Test</description>
+        </channel>
+      </rss>
+    XML
+
+    rss_with = SimpleRSS.parse(xml_with_percent)
+    rss_without = SimpleRSS.parse(xml_without_percent)
+
+    assert_equal rss_with.title.encoding, rss_without.title.encoding,
+                 "Encoding should be consistent regardless of '%' in content"
+    assert_equal Encoding::UTF_8, rss_without.title.encoding
+  end
+end

data/test/base/feed_attributes_test.rb

@@ -0,0 +1,26 @@
+require "test_helper"
+
+class FeedAttributesTest < Test::Unit::TestCase
+  def setup
+    # Add feed attribute tags before parsing
+    SimpleRSS.feed_tags << :"channel#custom:version"
+    SimpleRSS.feed_tags << :"feed#app:id"
+
+    @rss20 = SimpleRSS.parse open(File.dirname(__FILE__) + "/../data/rss20_with_channel_attrs.xml")
+    @atom = SimpleRSS.parse open(File.dirname(__FILE__) + "/../data/atom_with_feed_attrs.xml")
+  end
+
+  def teardown
+    # Clean up added tags
+    SimpleRSS.feed_tags.delete(:"channel#custom:version")
+    SimpleRSS.feed_tags.delete(:"feed#app:id")
+  end
+
+  def test_rss20_channel_attribute
+    assert_equal "2.0", @rss20.channel_custom_version
+  end
+
+  def test_atom_feed_attribute
+    assert_equal "12345", @atom.feed_app_id
+  end
+end

data/test/base/item_attributes_test.rb

@@ -0,0 +1,26 @@
+require "test_helper"
+
+class ItemAttributesTest < Test::Unit::TestCase
+  def setup
+    # Add item/entry attribute tags before parsing
+    SimpleRSS.item_tags << :"entry#custom:id"
+    SimpleRSS.item_tags << :"item#data-id"
+
+    @atom = SimpleRSS.parse open(File.dirname(__FILE__) + "/../data/atom_with_entry_attrs.xml")
+    @rss20 = SimpleRSS.parse open(File.dirname(__FILE__) + "/../data/rss20_with_item_attrs.xml")
+  end
+
+  def teardown
+    # Clean up added tags
+    SimpleRSS.item_tags.delete(:"entry#custom:id")
+    SimpleRSS.item_tags.delete(:"item#data-id")
+  end
+
+  def test_atom_entry_attribute
+    assert_equal "12345", @atom.entries.first[:entry_custom_id]
+  end
+
+  def test_rss20_item_attribute
+    assert_equal "67890", @rss20.items.first[:"item_data-id"]
+  end
+end

data/test/data/atom_with_entry_attrs.xml

@@ -0,0 +1,13 @@
+<?xml version="1.0" encoding="utf-8"?>
+<feed xmlns="http://www.w3.org/2005/Atom" xmlns:custom="http://example.org/custom">
+  <title type="text">Test Feed</title>
+  <updated>2005-07-31T12:29:29Z</updated>
+  <id>tag:example.org,2003:3</id>
+  <link rel="alternate" type="text/html" href="http://example.org/"/>
+  <entry custom:id="12345">
+    <title>Test Entry</title>
+    <link rel="alternate" type="text/html" href="http://example.org/entry/1"/>
+    <id>tag:example.org,2003:3.2397</id>
+    <updated>2005-07-31T12:29:29Z</updated>
+  </entry>
+</feed>

data/test/data/atom_with_feed_attrs.xml

@@ -0,0 +1,13 @@
+<?xml version="1.0" encoding="utf-8"?>
+<feed xmlns="http://www.w3.org/2005/Atom" xmlns:app="http://example.org/app" app:id="12345">
+  <title type="text">Test Feed</title>
+  <updated>2005-07-31T12:29:29Z</updated>
+  <id>tag:example.org,2003:3</id>
+  <link rel="alternate" type="text/html" href="http://example.org/"/>
+  <entry>
+    <title>Test Entry</title>
+    <link rel="alternate" type="text/html" href="http://example.org/entry/1"/>
+    <id>tag:example.org,2003:3.2397</id>
+    <updated>2005-07-31T12:29:29Z</updated>
+  </entry>
+</feed>