Merge branch 'bugfix/html-scrub-schemes' into 'develop'

lib/pleroma/html.ex: Fix scheme lists

See merge request pleroma/pleroma!377

config/config.exs

@@ -24,7 +24,23 @@ config :pleroma, Pleroma.Uploaders.S3,
 
 config :pleroma, :emoji, shortcode_globs: ["/emoji/custom/**/*.png"]
 
-config :pleroma, :uri_schemes, additionnal_schemes: []
+config :pleroma, :uri_schemes,
+  valid_schemes: [
+    "https",
+    "http",
+    "dat",
+    "dweb",
+    "gopher",
+    "ipfs",
+    "ipns",
+    "irc",
+    "ircs",
+    "magnet",
+    "mailto",
+    "mumble",
+    "ssb",
+    "xmpp"
+  ]
 
 # Configures the endpoint
 config :pleroma, Pleroma.Web.Endpoint,

lib/pleroma/formatter.ex

@@ -171,25 +171,8 @@ defmodule Pleroma.Formatter do
 
   @link_regex ~r/[0-9a-z+\-\.]+:[0-9a-z$-_.+!*'(),]+/ui
 
-  # IANA got a list https://www.iana.org/assignments/uri-schemes/ but
+  @uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
-  # Stuff like ipfs isn’t in it
+  @valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
-  # There is very niche stuff
-  @uri_schemes [
-    "https://",
-    "http://",
-    "dat://",
-    "dweb://",
-    "gopher://",
-    "ipfs://",
-    "ipns://",
-    "irc:",
-    "ircs:",
-    "magnet:",
-    "mailto:",
-    "mumble:",
-    "ssb://",
-    "xmpp:"
-  ]
 
   # TODO: make it use something other than @link_regex
   def html_escape(text, "text/html") do
@@ -207,14 +190,10 @@ 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 ++ additionnal_schemes) end)
+      |> Enum.filter(fn word -> String.starts_with?(word, @valid_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)

lib/pleroma/html.ex

@@ -36,10 +36,14 @@ defmodule Pleroma.HTML.Scrubber.TwitterText do
   paragraphs, breaks and links are allowed through the filter.
   """
 
+  @markup Application.get_env(:pleroma, :markup)
+  @uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
+  @valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
+
   require HtmlSanitizeEx.Scrubber.Meta
   alias HtmlSanitizeEx.Scrubber.Meta
 
-  @valid_schemes ["http", "https"]
+  alias Pleroma.HTML
 
   Meta.remove_cdata_sections_before_scrub()
   Meta.strip_comments()
@@ -56,11 +60,11 @@ defmodule Pleroma.HTML.Scrubber.TwitterText do
   Meta.allow_tag_with_these_attributes("span", [])
 
   # allow inline images for custom emoji
-  @markup Application.get_env(:pleroma, :markup)
   @allow_inline_images Keyword.get(@markup, :allow_inline_images)
 
   if @allow_inline_images do
-    Meta.allow_tag_with_uri_attributes("img", ["src"], @valid_schemes)
+    # restrict img tags to http/https only, because of MediaProxy.
+    Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
 
     Meta.allow_tag_with_these_attributes("img", [
       "width",
@@ -79,7 +83,11 @@ defmodule Pleroma.HTML.Scrubber.Default do
   require HtmlSanitizeEx.Scrubber.Meta
   alias HtmlSanitizeEx.Scrubber.Meta
 
-  @valid_schemes ["http", "https"]
+  alias Pleroma.HTML
+
+  @markup Application.get_env(:pleroma, :markup)
+  @uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
+  @valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
 
   Meta.remove_cdata_sections_before_scrub()
   Meta.strip_comments()
@@ -103,11 +111,11 @@ defmodule Pleroma.HTML.Scrubber.Default do
   Meta.allow_tag_with_these_attributes("u", [])
   Meta.allow_tag_with_these_attributes("ul", [])
 
-  @markup Application.get_env(:pleroma, :markup)
   @allow_inline_images Keyword.get(@markup, :allow_inline_images)
 
   if @allow_inline_images do
-    Meta.allow_tag_with_uri_attributes("img", ["src"], @valid_schemes)
+    # restrict img tags to http/https only, because of MediaProxy.
+    Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
 
     Meta.allow_tag_with_these_attributes("img", [
       "width",