add Recovery.Entities module

test/feed_reader/recovery/entities_test.exs

@@ -0,0 +1,40 @@
+defmodule FeedReader.Recovery.EntitiesTest do
+  use ExUnit.Case, async: true
+
+  alias FeedReader.Recovery.Entities
+
+  test "preserves valid named and numeric entities" do
+    xml = "& < > " ' © 😀"
+
+    assert normalize([xml]) == xml
+  end
+
+  test "escapes raw and invalid entity references" do
+    assert normalize(["A & B &unknown; C &not-an-entity D"]) ==
+             "A & B &unknown; C &not-an-entity D"
+  end
+
+  test "preserves ampersands inside CDATA" do
+    chunks = ["<![C", "DATA[<p>A & B</p>]]>"]
+
+    assert normalize(chunks) == "<![CDATA[<p>A & B</p>]]>"
+  end
+
+  test "preserves ampersands inside comments" do
+    chunks = ["<!-- source: A &", " B -->"]
+
+    assert normalize(chunks) == "<!-- source: A & B -->"
+  end
+
+  test "handles an entity split across chunks" do
+    assert normalize(["A &am", "p; B &#", "169; C"]) == "A &amp; B &#169; C"
+  end
+
+  test "handles CDATA and comment terminators split across chunks" do
+    chunks = ["<![CDATA[A & B]", "]><!-- C & D -", "->E & F"]
+
+    assert normalize(chunks) == "<![CDATA[A & B]]><!-- C & D -->E &amp; F"
+  end
+
+  defp normalize(chunks), do: chunks |> Entities.normalize() |> Enum.join()
+end