add Recovery.Attributes module

test/feed_reader/recovery/attributes_test.exs

@@ -0,0 +1,48 @@
+defmodule FeedReader.Recovery.AttributesTest do
+  use ExUnit.Case, async: true
+
+  alias FeedReader.Recovery.Attributes
+
+  test "preserves valid attributes" do
+    xml = ~s(<link href="https://example.com" rel="alternate" />)
+
+    assert normalize([xml]) == xml
+  end
+
+  test "repairs literal double quotes inside an attribute value" do
+    xml = ~s(<link title="How to "Sustain Heroku"" />)
+
+    assert normalize([xml]) == ~s(<link title="How to &quot;Sustain Heroku&quot;" />)
+  end
+
+  test "repairs malformed attributes split across chunks" do
+    chunks = [~s(<link title="How to "Sustain), ~s( Heroku"" />)]
+
+    assert normalize(chunks) == ~s(<link title="How to &quot;Sustain Heroku&quot;" />)
+  end
+
+  test "preserves empty attributes and single-quoted attributes" do
+    xml = ~s(<source url="" label='A "quoted" source' />)
+
+    assert normalize([xml]) == xml
+  end
+
+  test "does not repair quotes in CDATA or comments" do
+    xml = ~s(<![CDATA[<p title="A "quoted" value">Text</p>]]><!-- A "quoted" comment -->)
+
+    assert normalize([xml]) == xml
+  end
+
+  test "does not repair across adjacent elements" do
+    xml = ~s(<source url="">Source</source><category domain="region">News</category>)
+
+    assert normalize([xml]) == xml
+  end
+
+  test "preserves text outside markup" do
+    assert normalize(["before ", ~s(<item title="Story" />), " after"]) ==
+             ~s(before <item title="Story" /> after)
+  end
+
+  defp normalize(chunks), do: chunks |> Attributes.normalize() |> Enum.join()
+end