RichMedia: skip the quote fallback link when picking the card URL

Mastodon, Misskey and InlineQuotePolicy put "RE: <url>" in a .quote-inline
element at the start of a quote post. Its link was taken for the link card,
which then showed the quoted post a second time.

changelog.d/card-skip-quote-link.fix

@@ -0,0 +1 @@
+Link preview cards skip the fallback link of quote posts (`RE: <url>`), which pointed at the quoted post and showed it twice.

lib/pleroma/html.ex

@@ -66,14 +66,16 @@ defmodule Pleroma.HTML do
   end
 
   @spec extract_first_external_url_from_object(Pleroma.Object.t()) :: String.t() | nil
-  def extract_first_external_url_from_object(%{data: %{"content" => content}})
+  def extract_first_external_url_from_object(%{data: %{"content" => content} = data})
       when is_binary(content) do
+    # A quote post's fallback link ("RE: <url>", in a .quote-inline element or
+    # bare) points at the quoted post, which is shown already.
     content
     |> Floki.parse_fragment!()
+    |> Floki.filter_out(".quote-inline")
     |> Floki.find("a:not(.mention,.hashtag,.attachment,[rel~=\"tag\"])")
-    |> Enum.take(1)
     |> Floki.attribute("href")
-    |> Enum.at(0)
+    |> Enum.find(&(&1 != data["quoteUrl"]))
   end
 
   def extract_first_external_url_from_object(_), do: nil

test/pleroma/html_test.exs

@@ -299,5 +299,42 @@ defmodule Pleroma.HTMLTest do
 
       assert nil == HTML.extract_first_external_url_from_object(object)
     end
+
+    test "skips the link in a Mastodon quote fallback paragraph" do
+      object = %Object{
+        data: %{
+          "quoteUrl" => "https://raru.re/users/ocean/statuses/116235427788617987",
+          "content" =>
+            "<p class=\"quote-inline\">RE: <a href=\"https://raru.re/@ocean/116235427788617987\">raru.re/@ocean/116235427788617987</a></p>" <>
+              "<p>Berapa lapis?</p><p><a href=\"https://www.youtube.com/watch?v=S0dJMhWDRDE\">youtube.com/watch?v=S0dJMhWDRDE</a></p>"
+        }
+      }
+
+      assert HTML.extract_first_external_url_from_object(object) ==
+               "https://www.youtube.com/watch?v=S0dJMhWDRDE"
+    end
+
+    test "skips the link in a quote fallback span (Misskey, InlineQuotePolicy)" do
+      object = %Object{
+        data: %{
+          "content" =>
+            "hello<span class=\"quote-inline\"><br/><br/>RE: <a href=\"https://misskey.example/notes/abc\">https://misskey.example/notes/abc</a></span>"
+        }
+      }
+
+      assert nil == HTML.extract_first_external_url_from_object(object)
+    end
+
+    test "skips a link to the quoted post without fallback markup" do
+      object = %Object{
+        data: %{
+          "quoteUrl" => "https://remote.example/objects/1",
+          "content" =>
+            "RE: <a href=\"https://remote.example/objects/1\">https://remote.example/objects/1</a> <a href=\"https://cofe.example/\">cofe</a>"
+        }
+      }
+
+      assert HTML.extract_first_external_url_from_object(object) == "https://cofe.example/"
+    end
   end
 end