Rich Media: Do not cache URLs for preview statuses

Closes #1987

CHANGELOG.md

@@ -12,6 +12,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
 - Mastodon API: Search parameter `following` now correctly returns the followings rather than the followers
 - Mastodon API: Timelines hanging for (`number of posts with links * rich media timeout`) in the worst case.
   Reduced to just rich media timeout.
+- Mastodon API: Cards being wrong for preview statuses due to cache key collision
 - Password resets no longer processed for deactivated accounts
 
 ## [2.1.0] - 2020-08-28

lib/pleroma/html.ex

@@ -100,21 +100,27 @@ defp generate_scrubber_signature(scrubbers) do
     end)
   end
 
-  def extract_first_external_url(_, nil), do: {:error, "No content"}
+  def extract_first_external_url_from_object(%{data: %{"content" => content}} = object)
-
+      when is_binary(content) do
-  def extract_first_external_url(object, content) do
+    unless object.data["fake"] do
       key = "URL|#{object.id}"
 
       Cachex.fetch!(:scrubber_cache, key, fn _key ->
-      result =
+        {:commit, {:ok, extract_first_external_url(content)}}
+      end)
+    else
+      {:ok, extract_first_external_url(content)}
+    end
+  end
+
+  def extract_first_external_url_from_object(_), do: {:error, :no_content}
+
+  def extract_first_external_url(content) do
     content
     |> Floki.parse_fragment!()
     |> Floki.find("a:not(.mention,.hashtag,.attachment,[rel~=\"tag\"])")
     |> Enum.take(1)
     |> Floki.attribute("href")
     |> Enum.at(0)
-
-      {:commit, {:ok, result}}
-    end)
   end
 end

lib/pleroma/web/rich_media/helpers.ex

@@ -58,7 +58,7 @@ def fetch_data_for_object(object) do
     with true <- Config.get([:rich_media, :enabled]),
          false <- object.data["sensitive"] || false,
          {:ok, page_url} <-
-           HTML.extract_first_external_url(object, object.data["content"]),
+           HTML.extract_first_external_url_from_object(object),
          :ok <- validate_page_url(page_url),
          {:ok, rich_media} <- Parser.parse(page_url) do
       %{page_url: page_url, rich_media: rich_media}

test/html_test.exs

@@ -165,7 +165,7 @@ test "filters invalid microformats markup" do
     end
   end
 
-  describe "extract_first_external_url" do
+  describe "extract_first_external_url_from_object" do
     test "extracts the url" do
       user = insert(:user)
 
@@ -176,7 +176,7 @@ test "extracts the url" do
         })
 
       object = Object.normalize(activity)
-      {:ok, url} = HTML.extract_first_external_url(object, object.data["content"])
+      {:ok, url} = HTML.extract_first_external_url_from_object(object)
       assert url == "https://github.com/komeiji-satori/Dress"
     end
 
@@ -191,7 +191,7 @@ test "skips mentions" do
         })
 
       object = Object.normalize(activity)
-      {:ok, url} = HTML.extract_first_external_url(object, object.data["content"])
+      {:ok, url} = HTML.extract_first_external_url_from_object(object)
 
       assert url == "https://github.com/syuilo/misskey/blob/develop/docs/setup.en.md"
 
@@ -207,7 +207,7 @@ test "skips hashtags" do
         })
 
       object = Object.normalize(activity)
-      {:ok, url} = HTML.extract_first_external_url(object, object.data["content"])
+      {:ok, url} = HTML.extract_first_external_url_from_object(object)
 
       assert url == "https://www.pixiv.net/member_illust.php?mode=medium&illust_id=72255140"
     end
@@ -223,7 +223,7 @@ test "skips microformats hashtags" do
         })
 
       object = Object.normalize(activity)
-      {:ok, url} = HTML.extract_first_external_url(object, object.data["content"])
+      {:ok, url} = HTML.extract_first_external_url_from_object(object)
 
       assert url == "https://www.pixiv.net/member_illust.php?mode=medium&illust_id=72255140"
     end
@@ -235,7 +235,7 @@ test "does not crash when there is an HTML entity in a link" do
 
       object = Object.normalize(activity)
 
-      assert {:ok, nil} = HTML.extract_first_external_url(object, object.data["content"])
+      assert {:ok, nil} = HTML.extract_first_external_url_from_object(object)
     end
 
     test "skips attachment links" do
@@ -249,7 +249,7 @@ test "skips attachment links" do
 
       object = Object.normalize(activity)
 
-      assert {:ok, nil} = HTML.extract_first_external_url(object, object.data["content"])
+      assert {:ok, nil} = HTML.extract_first_external_url_from_object(object)
     end
   end
 end

test/web/mastodon_api/controllers/status_controller_test.exs

@@ -296,9 +296,45 @@ test "posting a fake status", %{conn: conn} do
       assert real_status == fake_status
     end
 
+    test "fake statuses' preview card is not cached", %{conn: conn} do
+      clear_config([:rich_media, :enabled], true)
+
+      Tesla.Mock.mock(fn
+        %{
+          method: :get,
+          url: "https://example.com/twitter-card"
+        } ->
+          %Tesla.Env{status: 200, body: File.read!("test/fixtures/rich_media/twitter_card.html")}
+
+        env ->
+          apply(HttpRequestMock, :request, [env])
+      end)
+
+      conn1 =
+        conn
+        |> put_req_header("content-type", "application/json")
+        |> post("/api/v1/statuses", %{
+          "status" => "https://example.com/ogp",
+          "preview" => true
+        })
+
+      conn2 =
+        conn
+        |> put_req_header("content-type", "application/json")
+        |> post("/api/v1/statuses", %{
+          "status" => "https://example.com/twitter-card",
+          "preview" => true
+        })
+
+      assert %{"card" => %{"title" => "The Rock"}} = json_response_and_validate_schema(conn1, 200)
+
+      assert %{"card" => %{"title" => "Small Island Developing States Photo Submission"}} =
+               json_response_and_validate_schema(conn2, 200)
+    end
+
     test "posting a status with OGP link preview", %{conn: conn} do
       Tesla.Mock.mock(fn env -> apply(HttpRequestMock, :request, [env]) end)
-      Config.put([:rich_media, :enabled], true)
+      clear_config([:rich_media, :enabled], true)
 
       conn =
         conn