Merge branch 'feature/html-scrub-policy' into 'develop'

html scrub policy

See merge request pleroma/pleroma!339

config/config.exs

@@ -76,6 +76,15 @@ config :pleroma, :instance,
   quarantined_instances: [],
   managed_config: true
 
+config :pleroma, :markup,
+  # XXX - unfortunately, inline images must be enabled by default right now, because
+  # of custom emoji.  Issue #275 discusses defanging that somehow.
+  allow_inline_images: true,
+  allow_headings: false,
+  allow_tables: false,
+  allow_fonts: false,
+  scrub_policy: Pleroma.HTML.Scrubber.Default
+
 config :pleroma, :fe,
   theme: "pleroma-dark",
   logo: "/static/logo.png",

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
@@ -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,
@@ -154,7 +155,7 @@ defmodule Pleroma.Formatter do
           MediaProxy.url(file)
         }' />"
       )
-      |> HtmlSanitizeEx.basic_html()
+      |> HTML.filter_tags()
     end)
   end
 

lib/pleroma/gopher/server.ex

@@ -35,6 +35,7 @@ defmodule Pleroma.Gopher.Server.ProtocolHandler do
   alias Pleroma.User
   alias Pleroma.Activity
   alias Pleroma.Repo
+  alias Pleroma.HTML
 
   @instance Application.get_env(:pleroma, :instance)
   @gopher Application.get_env(:pleroma, :gopher)
@@ -78,11 +79,7 @@ defmodule Pleroma.Gopher.Server.ProtocolHandler do
       link("Post ##{activity.id} by #{user.nickname}", "/notices/#{activity.id}") <>
         info("#{like_count} likes, #{announcement_count} repeats") <>
         "i\tfake\t(NULL)\t0\r\n" <>
-        info(
-          HtmlSanitizeEx.strip_tags(
-            String.replace(activity.data["object"]["content"], "<br>", "\r")
-          )
-        )
+        info(HTML.strip_tags(String.replace(activity.data["object"]["content"], "<br>", "\r")))
     end)
     |> Enum.join("i\tfake\t(NULL)\t0\r\n")
   end

lib/pleroma/html.ex

@@ -0,0 +1,129 @@
+defmodule Pleroma.HTML do
+  alias HtmlSanitizeEx.Scrubber
+
+  @markup Application.get_env(:pleroma, :markup)
+
+  def filter_tags(html) do
+    scrubber = Keyword.get(@markup, :scrub_policy)
+    html |> Scrubber.scrub(scrubber)
+  end
+
+  def strip_tags(html) do
+    html |> Scrubber.scrub(Scrubber.StripTags)
+  end
+end
+
+defmodule Pleroma.HTML.Scrubber.TwitterText do
+  @moduledoc """
+  An HTML scrubbing policy which limits to twitter-style text.  Only
+  paragraphs, breaks and links are allowed through the filter.
+  """
+
+  require HtmlSanitizeEx.Scrubber.Meta
+  alias HtmlSanitizeEx.Scrubber.Meta
+
+  @valid_schemes ["http", "https"]
+
+  Meta.remove_cdata_sections_before_scrub()
+  Meta.strip_comments()
+
+  # links
+  Meta.allow_tag_with_uri_attributes("a", ["href"], @valid_schemes)
+  Meta.allow_tag_with_these_attributes("a", ["name", "title"])
+
+  # paragraphs and linebreaks
+  Meta.allow_tag_with_these_attributes("br", [])
+  Meta.allow_tag_with_these_attributes("p", [])
+
+  # microformats
+  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)
+
+    Meta.allow_tag_with_these_attributes("img", [
+      "width",
+      "height",
+      "title",
+      "alt"
+    ])
+  end
+end
+
+defmodule Pleroma.HTML.Scrubber.Default do
+  @doc "The default HTML scrubbing policy: no "
+
+  require HtmlSanitizeEx.Scrubber.Meta
+  alias HtmlSanitizeEx.Scrubber.Meta
+
+  @valid_schemes ["http", "https"]
+
+  Meta.remove_cdata_sections_before_scrub()
+  Meta.strip_comments()
+
+  Meta.allow_tag_with_uri_attributes("a", ["href"], @valid_schemes)
+  Meta.allow_tag_with_these_attributes("a", ["name", "title"])
+
+  Meta.allow_tag_with_these_attributes("b", [])
+  Meta.allow_tag_with_these_attributes("blockquote", [])
+  Meta.allow_tag_with_these_attributes("br", [])
+  Meta.allow_tag_with_these_attributes("code", [])
+  Meta.allow_tag_with_these_attributes("del", [])
+  Meta.allow_tag_with_these_attributes("em", [])
+  Meta.allow_tag_with_these_attributes("i", [])
+  Meta.allow_tag_with_these_attributes("li", [])
+  Meta.allow_tag_with_these_attributes("ol", [])
+  Meta.allow_tag_with_these_attributes("p", [])
+  Meta.allow_tag_with_these_attributes("pre", [])
+  Meta.allow_tag_with_these_attributes("span", [])
+  Meta.allow_tag_with_these_attributes("strong", [])
+  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)
+
+    Meta.allow_tag_with_these_attributes("img", [
+      "width",
+      "height",
+      "title",
+      "alt"
+    ])
+  end
+
+  @allow_tables Keyword.get(@markup, :allow_tables)
+
+  if @allow_tables do
+    Meta.allow_tag_with_these_attributes("table", [])
+    Meta.allow_tag_with_these_attributes("tbody", [])
+    Meta.allow_tag_with_these_attributes("td", [])
+    Meta.allow_tag_with_these_attributes("th", [])
+    Meta.allow_tag_with_these_attributes("thead", [])
+    Meta.allow_tag_with_these_attributes("tr", [])
+  end
+
+  @allow_headings Keyword.get(@markup, :allow_headings)
+
+  if @allow_headings do
+    Meta.allow_tag_with_these_attributes("h1", [])
+    Meta.allow_tag_with_these_attributes("h2", [])
+    Meta.allow_tag_with_these_attributes("h3", [])
+    Meta.allow_tag_with_these_attributes("h4", [])
+    Meta.allow_tag_with_these_attributes("h5", [])
+  end
+
+  @allow_fonts Keyword.get(@markup, :allow_fonts)
+
+  if @allow_fonts do
+    Meta.allow_tag_with_these_attributes("font", ["face"])
+  end
+
+  Meta.strip_everything_not_covered()
+end

lib/pleroma/web/mastodon_api/views/account_view.ex

@@ -4,6 +4,7 @@ defmodule Pleroma.Web.MastodonAPI.AccountView do
   alias Pleroma.Web.MastodonAPI.AccountView
   alias Pleroma.Web.CommonAPI.Utils
   alias Pleroma.Web.MediaProxy
+  alias Pleroma.HTML
 
   def render("accounts.json", %{users: users} = opts) do
     render_many(users, AccountView, "account.json", opts)
@@ -42,7 +43,7 @@ defmodule Pleroma.Web.MastodonAPI.AccountView do
       followers_count: user_info.follower_count,
       following_count: user_info.following_count,
       statuses_count: user_info.note_count,
-      note: HtmlSanitizeEx.basic_html(user.bio) || "",
+      note: HTML.filter_tags(user.bio) || "",
       url: user.ap_id,
       avatar: image,
       avatar_static: image,

lib/pleroma/web/mastodon_api/views/status_view.ex

@@ -5,6 +5,7 @@ defmodule Pleroma.Web.MastodonAPI.StatusView do
   alias Pleroma.Web.CommonAPI.Utils
   alias Pleroma.Web.MediaProxy
   alias Pleroma.Repo
+  alias Pleroma.HTML
 
   # TODO: Add cached version.
   defp get_replied_to_activities(activities) do
@@ -111,10 +112,10 @@ defmodule Pleroma.Web.MastodonAPI.StatusView do
     emojis =
       (activity.data["object"]["emoji"] || [])
       |> Enum.map(fn {name, url} ->
-        name = HtmlSanitizeEx.strip_tags(name)
+        name = HTML.strip_tags(name)
 
         url =
-          HtmlSanitizeEx.strip_tags(url)
+          HTML.strip_tags(url)
           |> MediaProxy.url()
 
         %{shortcode: name, url: url, static_url: url, visible_in_picker: false}
@@ -221,7 +222,7 @@ defmodule Pleroma.Web.MastodonAPI.StatusView do
         object["content"]
       end
 
-    HtmlSanitizeEx.basic_html(content)
+    HTML.filter_tags(content)
   end
 
   def render_content(%{"type" => "Article"} = object) do
@@ -234,10 +235,10 @@ defmodule Pleroma.Web.MastodonAPI.StatusView do
         object["content"]
       end
 
-    HtmlSanitizeEx.basic_html(content)
+    HTML.filter_tags(content)
   end
 
   def render_content(object) do
-    HtmlSanitizeEx.basic_html(object["content"])
+    HTML.filter_tags(object["content"])
   end
 end

lib/pleroma/web/twitter_api/representers/activity_representer.ex

@@ -7,6 +7,7 @@ defmodule Pleroma.Web.TwitterAPI.Representers.ActivityRepresenter do
   alias Pleroma.Web.TwitterAPI.{TwitterAPI, UserView, ActivityView}
   alias Pleroma.Web.CommonAPI.Utils
   alias Pleroma.Formatter
+  alias Pleroma.HTML
 
   defp user_by_ap_id(user_list, ap_id) do
     Enum.find(user_list, fn %{ap_id: user_id} -> ap_id == user_id end)
@@ -167,7 +168,7 @@ defmodule Pleroma.Web.TwitterAPI.Representers.ActivityRepresenter do
     {summary, content} = ActivityView.render_content(object)
 
     html =
-      HtmlSanitizeEx.basic_html(content)
+      HTML.filter_tags(content)
       |> Formatter.emojify(object["emoji"])
 
     video =
@@ -184,7 +185,7 @@ defmodule Pleroma.Web.TwitterAPI.Representers.ActivityRepresenter do
       "uri" => activity.data["object"]["id"],
       "user" => UserView.render("show.json", %{user: user, for: opts[:for]}),
       "statusnet_html" => html,
-      "text" => HtmlSanitizeEx.strip_tags(content),
+      "text" => HTML.strip_tags(content),
       "is_local" => activity.local,
       "is_post_verb" => true,
       "created_at" => created_at,

lib/pleroma/web/twitter_api/views/activity_view.ex

@@ -11,6 +11,7 @@ defmodule Pleroma.Web.TwitterAPI.ActivityView do
   alias Pleroma.User
   alias Pleroma.Repo
   alias Pleroma.Formatter
+  alias Pleroma.HTML
 
   import Ecto.Query
 
@@ -232,7 +233,7 @@ defmodule Pleroma.Web.TwitterAPI.ActivityView do
     {summary, content} = render_content(object)
 
     html =
-      HtmlSanitizeEx.basic_html(content)
+      HTML.filter_tags(content)
       |> Formatter.emojify(object["emoji"])
 
     %{
@@ -240,7 +241,7 @@ defmodule Pleroma.Web.TwitterAPI.ActivityView do
       "uri" => activity.data["object"]["id"],
       "user" => UserView.render("show.json", %{user: user, for: opts[:for]}),
       "statusnet_html" => html,
-      "text" => HtmlSanitizeEx.strip_tags(content),
+      "text" => HTML.strip_tags(content),
       "is_local" => activity.local,
       "is_post_verb" => true,
       "created_at" => created_at,

lib/pleroma/web/twitter_api/views/user_view.ex

@@ -4,6 +4,7 @@ defmodule Pleroma.Web.TwitterAPI.UserView do
   alias Pleroma.Formatter
   alias Pleroma.Web.CommonAPI.Utils
   alias Pleroma.Web.MediaProxy
+  alias Pleroma.HTML
 
   def render("show.json", %{user: user = %User{}} = assigns) do
     render_one(user, Pleroma.Web.TwitterAPI.UserView, "user.json", assigns)
@@ -38,9 +39,8 @@ defmodule Pleroma.Web.TwitterAPI.UserView do
 
     data = %{
       "created_at" => user.inserted_at |> Utils.format_naive_asctime(),
-      "description" =>
-        HtmlSanitizeEx.strip_tags((user.bio || "") |> String.replace("<br>", "\n")),
-      "description_html" => HtmlSanitizeEx.basic_html(user.bio),
+      "description" => HTML.strip_tags((user.bio || "") |> String.replace("<br>", "\n")),
+      "description_html" => HTML.filter_tags(user.bio),
       "favourites_count" => 0,
       "followers_count" => user_info[:follower_count],
       "following" => following,
@@ -49,7 +49,7 @@ defmodule Pleroma.Web.TwitterAPI.UserView do
       "friends_count" => user_info[:following_count],
       "id" => user.id,
       "name" => user.name,
-      "name_html" => HtmlSanitizeEx.strip_tags(user.name) |> Formatter.emojify(emoji),
+      "name_html" => HTML.strip_tags(user.name) |> Formatter.emojify(emoji),
       "profile_image_url" => image,
       "profile_image_url_https" => image,
       "profile_image_url_profile_size" => image,