html: allow scrubbing policies to be stackable

[akkoma] / lib / pleroma / formatter.ex

diff --git a/lib/pleroma/formatter.ex b/lib/pleroma/formatter.ex
index fe3da09ac3e1e02481c6f3d1e76cf07e5961b86d..62f54a3f255e2e74d5e7c6dc7dfa7ff1f83d4e46 100644 (file)
--- a/lib/pleroma/formatter.ex
+++ b/lib/pleroma/formatter.ex
@@ -1,6 +1,7 @@
 defmodule Pleroma.Formatter do
   alias Pleroma.User
   alias Pleroma.Web.MediaProxy
+  alias Pleroma.HTML
 
   @tag_regex ~r/\#\w+/u
   def parse_tags(text, data \\ %{}) do
@@ -16,7 +17,7 @@ defmodule Pleroma.Formatter do
   def parse_mentions(text) do
     # Modified from https://www.w3.org/TR/html5/forms.html#valid-e-mail-address
     regex =
-      ~r/@[a-zA-Z0-9.!#$%&'*+\/=?^_`{|}~-]+@?[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*/u
+      ~r/@[a-zA-Z0-9.!#$%&'*+\/=?^_`{|}~-]*@?[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]{0,61}[a-zA-Z0-9])?)*/u
 
     Regex.scan(regex, text)
     |> List.flatten()
@@ -144,8 +145,8 @@ defmodule Pleroma.Formatter do
 
   def emojify(text, emoji) do
     Enum.reduce(emoji, text, fn {emoji, file}, text ->
-      emoji = HtmlSanitizeEx.strip_tags(emoji)
-      file = HtmlSanitizeEx.strip_tags(file)
+      emoji = HTML.strip_tags(emoji)
+      file = HTML.strip_tags(file)
 
       String.replace(
         text,
@@ -154,13 +155,16 @@ defmodule Pleroma.Formatter do
           MediaProxy.url(file)
         }' />"
       )
+      |> HTML.filter_tags()
     end)
   end
 
-  def get_emoji(text) do
+  def get_emoji(text) when is_binary(text) do
     Enum.filter(@emoji, fn {emoji, _} -> String.contains?(text, ":#{emoji}:") end)
   end
 
+  def get_emoji(_), do: []
+
   def get_custom_emoji() do
     @emoji
   end
@@ -199,10 +203,14 @@ defmodule Pleroma.Formatter do
 
   @doc "changes scheme:... urls to html links"
   def add_links({subs, text}) do
+    additionnal_schemes =
+      Application.get_env(:pleroma, :uri_schemes, [])
+      |> Keyword.get(:additionnal_schemes, [])
+
     links =
       text
       |> String.split([" ", "\t", "<br>"])
-      |> Enum.filter(fn word -> String.starts_with?(word, @uri_schemes) end)
+      |> Enum.filter(fn word -> String.starts_with?(word, @uri_schemes ++ additionnal_schemes) end)
       |> Enum.filter(fn word -> Regex.match?(@link_regex, word) end)
       |> Enum.map(fn url -> {Ecto.UUID.generate(), url} end)
       |> Enum.sort_by(fn {_, url} -> -String.length(url) end)
@@ -268,8 +276,8 @@ defmodule Pleroma.Formatter do
 
     subs =
       subs ++
-        Enum.map(tags, fn {_, tag, uuid} ->
-          url = "<a href='#{Pleroma.Web.base_url()}/tag/#{tag}' rel='tag'>##{tag}</a>"
+        Enum.map(tags, fn {tag_text, tag, uuid} ->
+          url = "<a href='#{Pleroma.Web.base_url()}/tag/#{tag}' rel='tag'>#{tag_text}</a>"
           {uuid, url}
         end)