better formatting for titles

config/config.exs

@@ -286,7 +286,7 @@ config :pleroma, :feed,
     max_length: 100,
     omission: "...",
     # New method to extract title
-    parse_source: true
+    parse_source: false
   }
 
 config :pleroma, :markup,

config/description.exs

@@ -2919,6 +2919,7 @@ config :pleroma, :config_description, [
   %{
     group: :pleroma,
     key: :feed,
+    label: "RSS Feeds",
     type: :group,
     description: "Configure feed rendering",
     children: [

lib/pleroma/formatter.ex

@@ -156,10 +156,13 @@ defmodule Pleroma.Formatter do
 
     length_with_omission = max_length - String.length(omission)
 
-    if length_with_omission <= 0 do
-      String.slice(text, 0, max_length)
-    else
-      String.slice(text, 0, length_with_omission) <> omission
+    cond do
+      String.length(text) <= max_length ->
+        text
+      length_with_omission > 0 ->
+        String.slice(text, 0, length_with_omission) <> omission
+      true ->
+        String.slice(text, 0, max_length)
     end
   end
 

lib/pleroma/web/feed/feed_view.ex

@@ -86,6 +86,7 @@ defmodule Pleroma.Web.Feed.FeedView do
     split_content(content, content_type, opts) |> elem(0)
   end
 
+  # TODO: scrub_html should replace <p> with " "
   def activity_title(%{"content" => content}, opts) do
     content
     |> Pleroma.Web.Metadata.Utils.scrub_html()

lib/pleroma/web/metadata/utils.ex

@@ -42,8 +42,12 @@ defmodule Pleroma.Web.Metadata.Utils do
     content
     # html content comes from DB already encoded, decode first and scrub after
     |> HtmlEntities.decode()
-    |> String.replace(~r/<br\s?\/?>/, " ")
+    |> String.replace(~r/<(br|p)[^>]*>/, "\\0&nbsp;")
+    |> String.replace(~r/<\/p\s*\/?>/, "&nbsp;\\0")
     |> HTML.strip_tags()
+    # strip_tags will convert &nbsp; to U+00A0, adding /u will match these to " "
+    |> String.replace(~r/\s+/u, " ")
+    |> String.trim()
   end
 
   def scrub_html(content), do: content