Merge pull request 'Add htmlMfm key when relevant' (#878) from ilja/akkoma:add_fep-c16b_discovery_mechanism_to_not_always_reparse_mfm into develop

Reviewed-on: https://akkoma.dev/AkkomaGang/akkoma/pulls/878
Reviewed-by: Oneric <oneric@noreply.akkoma>

CHANGELOG.md

@@ -7,6 +7,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
 ## Unreleased
 
 ### Added
+- We mark our MFM posts as FEP-c16b compliant, and retain remote HTML representations for incoming posts marked as FEP-c16b-compliant. (Safety scrubbers are still applied)
 - Prometheus stats now exposes failed ActivityPub deliveries
   which failed all attempts and the failure reason
 - status and user HTML pages now provide ActivityPub alternate links

FEDERATION.md

@@ -34,7 +34,8 @@ Depending on instance configuration the same may be true for GET requests.
 
 ### FEP-c16b: Formatting MFM functions
 
-The optional extension term `htmlMfm` is currently not used.
+We set the optional extension term `htmlMfm: true` when using content type "text/x.misskeymarkdown".
+Incoming messages containing `htmlMfm: true` will not have their content re-parsed.
 
 ## Nodeinfo
 

lib/pleroma/web/activity_pub/object_validators/article_note_page_validator.ex

@@ -124,6 +124,8 @@ defp remote_mention_resolver(
     end
   end
 
+  defp fix_misskey_content(object = %{"htmlMfm" => true}), do: object
+
   # See https://akkoma.dev/FoundKeyGang/FoundKey/issues/343
   # Misskey/Foundkey drops some of the custom formatting when it sends remotely
   # So this basically reprocesses the MFM source
@@ -144,20 +146,13 @@ defp fix_misskey_content(
   # See https://github.com/misskey-dev/misskey/pull/8787
   # This is for compatibility with older Misskey instances
   defp fix_misskey_content(%{"_misskey_content" => content} = object) when is_binary(content) do
-    mention_handler = fn nick, buffer, opts, acc ->
-      remote_mention_resolver(object, nick, buffer, opts, acc)
-    end
-
-    {linked, _, _} =
-      Utils.format_input(content, "text/x.misskeymarkdown", mention_handler: mention_handler)
-
     object
     |> Map.put("source", %{
       "content" => content,
       "mediaType" => "text/x.misskeymarkdown"
     })
-    |> Map.put("content", linked)
     |> Map.delete("_misskey_content")
+    |> fix_misskey_content()
   end
 
   defp fix_misskey_content(data), do: data

lib/pleroma/web/activity_pub/object_validators/common_fields.ex

@@ -31,6 +31,7 @@ defmacro activity_fields do
   defmacro object_fields do
     quote bind_quoted: binding() do
       field(:content, :string)
+      field(:htmlMfm, :boolean)
 
       field(:published, ObjectValidators.DateTime)
       field(:updated, ObjectValidators.DateTime)

lib/pleroma/web/activity_pub/utils.ex

@@ -102,7 +102,8 @@ def make_json_ld_header do
         "https://www.w3.org/ns/activitystreams",
         "#{Endpoint.url()}/schemas/litepub-0.1.jsonld",
         %{
-          "@language" => "und"
+          "@language" => "und",
+          "htmlMfm" => "https://w3id.org/fep/c16b#htmlMfm"
         }
       ]
     }

lib/pleroma/web/common_api/activity_draft.ex

@@ -47,8 +47,7 @@ defp new(user, params) do
   end
 
   def create(user, params) do
-    user
-    |> new(params)
+    new(user, params)
     |> status()
     |> summary()
     |> with_valid(&attachments/1)
@@ -236,6 +235,7 @@ defp object(draft) do
       end
 
     emoji = Map.merge(emoji, summary_emoji)
+    media_type = Utils.get_content_type(draft.params[:content_type])
     {:ok, note_data, _meta} = Builder.note(draft)
 
     object =
@@ -243,14 +243,18 @@ defp object(draft) do
       |> Map.put("emoji", emoji)
       |> Map.put("source", %{
         "content" => draft.status,
-        "mediaType" => Utils.get_content_type(draft.params[:content_type])
+        "mediaType" => media_type
       })
+      |> maybe_put("htmlMfm", true, media_type == "text/x.misskeymarkdown")
       |> Map.put("generator", draft.params[:generator])
       |> Map.put("contentMap", draft.content_map)
 
     %__MODULE__{draft | object: object}
   end
 
+  defp maybe_put(map, key, value, true), do: map |> Map.put(key, value)
+  defp maybe_put(map, _, _, _), do: map
+
   defp preview?(draft) do
     preview? = Pleroma.Web.Utils.Params.truthy_param?(draft.params[:preview])
     %__MODULE__{draft | preview?: preview?}

test/pleroma/web/activity_pub/object_validators/article_note_page_validator_test.exs

@@ -191,6 +191,31 @@ test "a misskey MFM status with a _misskey_content field should work and be link
     end
   end
 
+  test "an MFM status with htmlMfm should not have its content re-parsed" do
+    insert(:user, %{ap_id: "https://misskey.local.live/users/92hzkskwgy"})
+
+    note =
+      "test/fixtures/misskey/mfm_x_format.json"
+      |> File.read!()
+      |> Jason.decode!()
+      |> Map.put("htmlMfm", true)
+      |> Map.put("content", "do not re-parse")
+
+    changes = ArticleNotePageValidator.cast_and_validate(note)
+
+    %{
+      valid?: true,
+      changes: %{
+        content: content,
+        source: %{
+          "mediaType" => "text/x.misskeymarkdown"
+        }
+      }
+    } = changes
+
+    assert content == "do not re-parse"
+  end
+
   test "a Note without replies/first/items validates" do
     insert(:user, ap_id: "https://mastodon.social/users/emelie")
 

test/pleroma/web/activity_pub/utils_test.exs

@@ -144,7 +144,8 @@ test "make_json_ld_header/0" do
                "https://www.w3.org/ns/activitystreams",
                "http://localhost:4001/schemas/litepub-0.1.jsonld",
                %{
-                 "@language" => "und"
+                 "@language" => "und",
+                 "htmlMfm" => "https://w3id.org/fep/c16b#htmlMfm"
                }
              ]
            }

test/pleroma/web/common_api_test.exs

@@ -534,6 +534,34 @@ test "it allows to address a list" do
       assert activity.data["listMessage"] == list.ap_id
     end
 
+    test "it adds the htmlMFM term to MFM posts and properly processes it" do
+      user = insert(:user)
+
+      assert {:ok,
+              %Pleroma.Activity{
+                object: %Pleroma.Object{
+                  data: %{
+                    "content" => content,
+                    "source" => %{
+                      "content" => source_content,
+                      "mediaType" => "text/x.misskeymarkdown"
+                    },
+                    "htmlMfm" => html_mfm
+                  }
+                }
+              }} =
+               CommonAPI.post(user, %{
+                 status: "<p class='scrub-this'>$[spin 13:37]</p>",
+                 content_type: "text/x.misskeymarkdown"
+               })
+
+      assert html_mfm == true
+      assert content =~ "mfm-spin"
+      assert content =~ "13:37"
+      refute content =~ "scrub-this"
+      assert source_content == "<p class='scrub-this'>$[spin 13:37]</p>"
+    end
+
     test "it returns error when status is empty and no attachments" do
       user = insert(:user)
 

test/pleroma/web/federator_test.exs

@@ -135,6 +135,105 @@ test "successfully processes incoming AP docs with correct origin" do
       assert {:cancel, :already_present} = ObanHelpers.perform(job)
     end
 
+    test "properly processes objects with the htmlMfm attribute true" do
+      params = %{
+        "@context" => "https://www.w3.org/ns/activitystreams",
+        "actor" => "http://mastodon.example.org/users/admin",
+        "type" => "Create",
+        "id" => "http://mastodon.example.org/users/admin/activities/1",
+        "object" => %{
+          "type" => "Note",
+          "content" => "<p this-should-be-scrubbed-away>this is the original content</p>",
+          "source" => %{
+            "content" =>
+              "this source content is irrelevant because the object content should be kept",
+            "mediaType" => "text/x.misskeymarkdown"
+          },
+          "htmlMfm" => true,
+          "id" => "http://mastodon.example.org/users/admin/objects/1",
+          "attributedTo" => "http://mastodon.example.org/users/admin",
+          "to" => ["https://www.w3.org/ns/activitystreams#Public"]
+        },
+        "to" => ["https://www.w3.org/ns/activitystreams#Public"]
+      }
+
+      {:ok, job} = Federator.incoming_ap_doc(params)
+      {:ok, %Pleroma.Activity{data: %{"object" => object_ap_id}}} = ObanHelpers.perform(job)
+
+      %Pleroma.Object{data: %{"content" => content, "htmlMfm" => html_mfm}} =
+        Pleroma.Object.get_by_ap_id(object_ap_id)
+
+      assert html_mfm == true
+      refute content =~ "this-should-be-scrubbed-away"
+      refute content =~ "source content is irrelevant"
+      assert content =~ "this is the original content"
+    end
+
+    test "properly processes objects with the htmlMfm attribute false" do
+      params = %{
+        "@context" => "https://www.w3.org/ns/activitystreams",
+        "actor" => "http://mastodon.example.org/users/admin",
+        "type" => "Create",
+        "id" => "http://mastodon.example.org/users/admin/activities/1",
+        "object" => %{
+          "type" => "Note",
+          "content" => "<p this-should-be-scrubbed-away>this is the original content</p>",
+          "source" => %{
+            "content" => "<p this-should-be-scrubbed-away>$[spin the source content is used]</p>",
+            "mediaType" => "text/x.misskeymarkdown"
+          },
+          "htmlMfm" => false,
+          "id" => "http://mastodon.example.org/users/admin/objects/1",
+          "attributedTo" => "http://mastodon.example.org/users/admin",
+          "to" => ["https://www.w3.org/ns/activitystreams#Public"]
+        },
+        "to" => ["https://www.w3.org/ns/activitystreams#Public"]
+      }
+
+      {:ok, job} = Federator.incoming_ap_doc(params)
+      {:ok, %Pleroma.Activity{data: %{"object" => object_ap_id}}} = ObanHelpers.perform(job)
+
+      %Pleroma.Object{data: %{"content" => content, "htmlMfm" => html_mfm}} =
+        Pleroma.Object.get_by_ap_id(object_ap_id)
+
+      assert html_mfm == false
+      refute content =~ "this-should-be-scrubbed-away"
+      assert content =~ "the source content is used"
+      refute content =~ "this is the original content"
+    end
+
+    test "properly processes objects with the htmlMfm attribute not set" do
+      params = %{
+        "@context" => "https://www.w3.org/ns/activitystreams",
+        "actor" => "http://mastodon.example.org/users/admin",
+        "type" => "Create",
+        "id" => "http://mastodon.example.org/users/admin/activities/1",
+        "object" => %{
+          "type" => "Note",
+          "content" => "<p this-should-be-scrubbed-away>this is the original content</p>",
+          "source" => %{
+            "content" => "<p this-should-be-scrubbed-away>$[spin the source content is used]</p>",
+            "mediaType" => "text/x.misskeymarkdown"
+          },
+          "id" => "http://mastodon.example.org/users/admin/objects/1",
+          "attributedTo" => "http://mastodon.example.org/users/admin",
+          "to" => ["https://www.w3.org/ns/activitystreams#Public"]
+        },
+        "to" => ["https://www.w3.org/ns/activitystreams#Public"]
+      }
+
+      {:ok, job} = Federator.incoming_ap_doc(params)
+      {:ok, %Pleroma.Activity{data: %{"object" => object_ap_id}}} = ObanHelpers.perform(job)
+
+      %Pleroma.Object{data: %{"content" => content} = data} =
+        Pleroma.Object.get_by_ap_id(object_ap_id)
+
+      assert Map.get(data, "htmlMfm") == nil
+      refute content =~ "this-should-be-scrubbed-away"
+      assert content =~ "the source content is used"
+      refute content =~ "this is the original content"
+    end
+
     test "successfully normalises public scope descriptors" do
       params = %{
         "@context" => "https://www.w3.org/ns/activitystreams",