rip out fetch_initial_posts

Every time someone tries to use it, it goes mad and tries to scrape the
entire fediverse for no visible reason, it's better to just remove it
than continue shipping it in it's current state.

idea acked by lain and feld on irc

Closes #1595 #1422

config/config.exs

@@ -504,10 +504,6 @@ config :pleroma, :workers,
     federator_outgoing: 5
   ]
 
-config :pleroma, :fetch_initial_posts,
-  enabled: false,
-  pages: 5
-
 config :auto_linker,
   opts: [
     extra: true,

config/description.exs

@@ -2008,25 +2008,6 @@ config :pleroma, :config_description, [
     ]
   },
   %{
-    group: :pleroma,
-    key: :fetch_initial_posts,
-    type: :group,
-    description: "Fetching initial posts settings",
-    children: [
-      %{
-        key: :enabled,
-        type: :boolean,
-        description: "Fetch posts when a new user is federated with"
-      },
-      %{
-        key: :pages,
-        type: :integer,
-        description: "The amount of pages to fetch",
-        suggestions: [5]
-      }
-    ]
-  },
-  %{
     group: :auto_linker,
     key: :opts,
     type: :group,

docs/configuration/cheatsheet.md

@@ -151,14 +151,6 @@ config :pleroma, :mrf_user_allowlist,
 * `sign_object_fetches`: Sign object fetches with HTTP signatures
 * `authorized_fetch_mode`: Require HTTP signatures for AP fetches
 
-### :fetch_initial_posts
-
-!!! warning
-    Be careful with this setting, fetching posts may lead to new users being discovered whose posts will then also be fetched. This can lead to serious load on your instance and database.
-
-* `enabled`: If enabled, when a new user is discovered by your instance, fetch some of their latest posts.
-* `pages`: The amount of pages to fetch
-
 ## Pleroma.ScheduledActivity
 
 * `daily_user_limit`: the number of scheduled activities a user is allowed to create in a single day (Default: `25`)

lib/pleroma/user.ex

@@ -839,10 +839,6 @@ defmodule Pleroma.User do
       _e ->
         with [_nick, _domain] <- String.split(nickname, "@"),
              {:ok, user} <- fetch_by_nickname(nickname) do
-          if Pleroma.Config.get([:fetch_initial_posts, :enabled]) do
-            fetch_initial_posts(user)
-          end
-
           {:ok, user}
         else
           _e -> {:error, "not found " <> nickname}
@@ -850,11 +846,6 @@ defmodule Pleroma.User do
     end
   end
 
-  @doc "Fetch some posts when the user has just been federated with"
-  def fetch_initial_posts(user) do
-    BackgroundWorker.enqueue("fetch_initial_posts", %{"user_id" => user.id})
-  end
-
   @spec get_followers_query(User.t(), pos_integer() | nil) :: Ecto.Query.t()
   def get_followers_query(%User{} = user, nil) do
     User.Query.build(%{followers: user, deactivated: false})
@@ -1320,16 +1311,6 @@ defmodule Pleroma.User do
     Repo.delete(user)
   end
 
-  def perform(:fetch_initial_posts, %User{} = user) do
-    pages = Pleroma.Config.get!([:fetch_initial_posts, :pages])
-
-    # Insert all the posts in reverse order, so they're in the right order on the timeline
-    user.source_data["outbox"]
-    |> Utils.fetch_ordered_collection(pages)
-    |> Enum.reverse()
-    |> Enum.each(&Pleroma.Web.Federator.incoming_ap_doc/1)
-  end
-
   def perform(:deactivate_async, user, status), do: deactivate(user, status)
 
   @spec perform(atom(), User.t(), list()) :: list() | {:error, any()}
@@ -1458,18 +1439,7 @@ defmodule Pleroma.User do
     if !is_nil(user) and !needs_update?(user) do
       {:ok, user}
     else
-      # Whether to fetch initial posts for the user (if it's a new user & the fetching is enabled)
+      fetch_by_ap_id(ap_id)
-      should_fetch_initial = is_nil(user) and Pleroma.Config.get([:fetch_initial_posts, :enabled])
-
-      resp = fetch_by_ap_id(ap_id)
-
-      if should_fetch_initial do
-        with {:ok, %User{} = user} <- resp do
-          fetch_initial_posts(user)
-        end
-      end
-
-      resp
     end
   end
 

lib/pleroma/web/activity_pub/utils.ex

@@ -784,45 +784,6 @@ defmodule Pleroma.Web.ActivityPub.Utils do
 
   defp build_flag_object(_), do: []
 
-  @doc """
-  Fetches the OrderedCollection/OrderedCollectionPage from `from`, limiting the amount of pages fetched after
-  the first one to `pages_left` pages.
-  If the amount of pages is higher than the collection has, it returns whatever was there.
-  """
-  def fetch_ordered_collection(from, pages_left, acc \\ []) do
-    with {:ok, response} <- Tesla.get(from),
-         {:ok, collection} <- Jason.decode(response.body) do
-      case collection["type"] do
-        "OrderedCollection" ->
-          # If we've encountered the OrderedCollection and not the page,
-          # just call the same function on the page address
-          fetch_ordered_collection(collection["first"], pages_left)
-
-        "OrderedCollectionPage" ->
-          if pages_left > 0 do
-            # There are still more pages
-            if Map.has_key?(collection, "next") do
-              # There are still more pages, go deeper saving what we have into the accumulator
-              fetch_ordered_collection(
-                collection["next"],
-                pages_left - 1,
-                acc ++ collection["orderedItems"]
-              )
-            else
-              # No more pages left, just return whatever we already have
-              acc ++ collection["orderedItems"]
-            end
-          else
-            # Got the amount of pages needed, add them all to the accumulator
-            acc ++ collection["orderedItems"]
-          end
-
-        _ ->
-          {:error, "Not an OrderedCollection or OrderedCollectionPage"}
-      end
-    end
-  end
-
   #### Report-related helpers
   def get_reports(params, page, page_size) do
     params =

lib/pleroma/workers/background_worker.ex

@@ -10,10 +10,6 @@ defmodule Pleroma.Workers.BackgroundWorker do
   use Pleroma.Workers.WorkerHelper, queue: "background"
 
   @impl Oban.Worker
-  def perform(%{"op" => "fetch_initial_posts", "user_id" => user_id}, _job) do
-    user = User.get_cached_by_id(user_id)
-    User.perform(:fetch_initial_posts, user)
-  end
 
   def perform(%{"op" => "deactivate_user", "user_id" => user_id, "status" => status}, _job) do
     user = User.get_cached_by_id(user_id)

priv/repo/migrations/20200314123607_config_remove_fetch_initial_posts.exs

@@ -0,0 +1,10 @@
+defmodule Pleroma.Repo.Migrations.ConfigRemoveFetchInitialPosts do
+  use Ecto.Migration
+
+  def change do
+    execute(
+      "delete from config where config.key = ':fetch_initial_posts' and config.group = ':pleroma';",
+      ""
+    )
+  end
+end

priv/repo/migrations/20200315125756_delete_fetch_initial_posts_jobs.exs

@@ -0,0 +1,10 @@
+defmodule Pleroma.Repo.Migrations.DeleteFetchInitialPostsJobs do
+  use Ecto.Migration
+
+  def change do
+    execute(
+      "delete from oban_jobs where worker = 'Pleroma.Workers.BackgroundWorker' and args->>'op' = 'fetch_initial_posts';",
+      ""
+    )
+  end
+end

test/web/activity_pub/utils_test.exs

@@ -177,71 +177,6 @@ defmodule Pleroma.Web.ActivityPub.UtilsTest do
     end
   end
 
-  describe "fetch_ordered_collection" do
-    import Tesla.Mock
-
-    test "fetches the first OrderedCollectionPage when an OrderedCollection is encountered" do
-      mock(fn
-        %{method: :get, url: "http://mastodon.com/outbox"} ->
-          json(%{"type" => "OrderedCollection", "first" => "http://mastodon.com/outbox?page=true"})
-
-        %{method: :get, url: "http://mastodon.com/outbox?page=true"} ->
-          json(%{"type" => "OrderedCollectionPage", "orderedItems" => ["ok"]})
-      end)
-
-      assert Utils.fetch_ordered_collection("http://mastodon.com/outbox", 1) == ["ok"]
-    end
-
-    test "fetches several pages in the right order one after another, but only the specified amount" do
-      mock(fn
-        %{method: :get, url: "http://example.com/outbox"} ->
-          json(%{
-            "type" => "OrderedCollectionPage",
-            "orderedItems" => [0],
-            "next" => "http://example.com/outbox?page=1"
-          })
-
-        %{method: :get, url: "http://example.com/outbox?page=1"} ->
-          json(%{
-            "type" => "OrderedCollectionPage",
-            "orderedItems" => [1],
-            "next" => "http://example.com/outbox?page=2"
-          })
-
-        %{method: :get, url: "http://example.com/outbox?page=2"} ->
-          json(%{"type" => "OrderedCollectionPage", "orderedItems" => [2]})
-      end)
-
-      assert Utils.fetch_ordered_collection("http://example.com/outbox", 0) == [0]
-      assert Utils.fetch_ordered_collection("http://example.com/outbox", 1) == [0, 1]
-    end
-
-    test "returns an error if the url doesn't have an OrderedCollection/Page" do
-      mock(fn
-        %{method: :get, url: "http://example.com/not-an-outbox"} ->
-          json(%{"type" => "NotAnOutbox"})
-      end)
-
-      assert {:error, _} = Utils.fetch_ordered_collection("http://example.com/not-an-outbox", 1)
-    end
-
-    test "returns the what was collected if there are less pages than specified" do
-      mock(fn
-        %{method: :get, url: "http://example.com/outbox"} ->
-          json(%{
-            "type" => "OrderedCollectionPage",
-            "orderedItems" => [0],
-            "next" => "http://example.com/outbox?page=1"
-          })
-
-        %{method: :get, url: "http://example.com/outbox?page=1"} ->
-          json(%{"type" => "OrderedCollectionPage", "orderedItems" => [1]})
-      end)
-
-      assert Utils.fetch_ordered_collection("http://example.com/outbox", 5) == [0, 1]
-    end
-  end
-
   test "make_json_ld_header/0" do
     assert Utils.make_json_ld_header() == %{
              "@context" => [