git.squeep.com Git - akkoma/blob - lib/pleroma/html.ex

   1 # Pleroma: A lightweight social networking server
   2 # Copyright © 2017-2019 Pleroma Authors <https://pleroma.social/>
   3 # SPDX-License-Identifier: AGPL-3.0-only
   4
   5 defmodule Pleroma.HTML do
   6   alias HtmlSanitizeEx.Scrubber
   7
   8   defp get_scrubbers(scrubber) when is_atom(scrubber), do: [scrubber]
   9   defp get_scrubbers(scrubbers) when is_list(scrubbers), do: scrubbers
  10   defp get_scrubbers(_), do: [Pleroma.HTML.Scrubber.Default]
  11
  12   def get_scrubbers() do
  13     Pleroma.Config.get([:markup, :scrub_policy])
  14     |> get_scrubbers
  15   end
  16
  17   def filter_tags(html, nil) do
  18     filter_tags(html, get_scrubbers())
  19   end
  20
  21   def filter_tags(html, scrubbers) when is_list(scrubbers) do
  22     Enum.reduce(scrubbers, html, fn scrubber, html ->
  23       filter_tags(html, scrubber)
  24     end)
  25   end
  26
  27   def filter_tags(html, scrubber), do: Scrubber.scrub(html, scrubber)
  28   def filter_tags(html), do: filter_tags(html, nil)
  29   def strip_tags(html), do: Scrubber.scrub(html, Scrubber.StripTags)
  30
  31   def get_cached_scrubbed_html_for_object(content, scrubbers, object, module) do
  32     key = "#{module}#{generate_scrubber_signature(scrubbers)}|#{object.id}"
  33     Cachex.fetch!(:scrubber_cache, key, fn _key -> ensure_scrubbed_html(content, scrubbers) end)
  34   end
  35
  36   def get_cached_stripped_html_for_object(content, object, module) do
  37     get_cached_scrubbed_html_for_object(
  38       content,
  39       HtmlSanitizeEx.Scrubber.StripTags,
  40       object,
  41       module
  42     )
  43   end
  44
  45   def ensure_scrubbed_html(
  46         content,
  47         scrubbers
  48       ) do
  49     {:commit, filter_tags(content, scrubbers)}
  50   end
  51
  52   defp generate_scrubber_signature(scrubber) when is_atom(scrubber) do
  53     generate_scrubber_signature([scrubber])
  54   end
  55
  56   defp generate_scrubber_signature(scrubbers) do
  57     Enum.reduce(scrubbers, "", fn scrubber, signature ->
  58       "#{signature}#{to_string(scrubber)}"
  59     end)
  60   end
  61 end
  62
  63 defmodule Pleroma.HTML.Scrubber.TwitterText do
  64   @moduledoc """
  65   An HTML scrubbing policy which limits to twitter-style text.  Only
  66   paragraphs, breaks and links are allowed through the filter.
  67   """
  68
  69   @markup Application.get_env(:pleroma, :markup)
  70   @uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
  71   @valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
  72
  73   require HtmlSanitizeEx.Scrubber.Meta
  74   alias HtmlSanitizeEx.Scrubber.Meta
  75
  76   Meta.remove_cdata_sections_before_scrub()
  77   Meta.strip_comments()
  78
  79   # links
  80   Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
  81   Meta.allow_tag_with_these_attributes("a", ["name", "title", "class"])
  82
  83   # paragraphs and linebreaks
  84   Meta.allow_tag_with_these_attributes("br", [])
  85   Meta.allow_tag_with_these_attributes("p", [])
  86
  87   # microformats
  88   Meta.allow_tag_with_these_attributes("span", ["class"])
  89
  90   # allow inline images for custom emoji
  91   @allow_inline_images Keyword.get(@markup, :allow_inline_images)
  92
  93   if @allow_inline_images do
  94     # restrict img tags to http/https only, because of MediaProxy.
  95     Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
  96
  97     Meta.allow_tag_with_these_attributes("img", [
  98       "width",
  99       "height",
 100       "title",
 101       "alt"
 102     ])
 103   end
 104
 105   Meta.strip_everything_not_covered()
 106 end
 107
 108 defmodule Pleroma.HTML.Scrubber.Default do
 109   @doc "The default HTML scrubbing policy: no "
 110
 111   require HtmlSanitizeEx.Scrubber.Meta
 112   alias HtmlSanitizeEx.Scrubber.Meta
 113
 114   @markup Application.get_env(:pleroma, :markup)
 115   @uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
 116   @valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
 117
 118   Meta.remove_cdata_sections_before_scrub()
 119   Meta.strip_comments()
 120
 121   Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
 122   Meta.allow_tag_with_these_attributes("a", ["name", "title", "class"])
 123
 124   Meta.allow_tag_with_these_attributes("abbr", ["title"])
 125
 126   Meta.allow_tag_with_these_attributes("b", [])
 127   Meta.allow_tag_with_these_attributes("blockquote", [])
 128   Meta.allow_tag_with_these_attributes("br", [])
 129   Meta.allow_tag_with_these_attributes("code", [])
 130   Meta.allow_tag_with_these_attributes("del", [])
 131   Meta.allow_tag_with_these_attributes("em", [])
 132   Meta.allow_tag_with_these_attributes("i", [])
 133   Meta.allow_tag_with_these_attributes("li", [])
 134   Meta.allow_tag_with_these_attributes("ol", [])
 135   Meta.allow_tag_with_these_attributes("p", [])
 136   Meta.allow_tag_with_these_attributes("pre", [])
 137   Meta.allow_tag_with_these_attributes("span", ["class"])
 138   Meta.allow_tag_with_these_attributes("strong", [])
 139   Meta.allow_tag_with_these_attributes("u", [])
 140   Meta.allow_tag_with_these_attributes("ul", [])
 141
 142   @allow_inline_images Keyword.get(@markup, :allow_inline_images)
 143
 144   if @allow_inline_images do
 145     # restrict img tags to http/https only, because of MediaProxy.
 146     Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
 147
 148     Meta.allow_tag_with_these_attributes("img", [
 149       "width",
 150       "height",
 151       "title",
 152       "alt"
 153     ])
 154   end
 155
 156   @allow_tables Keyword.get(@markup, :allow_tables)
 157
 158   if @allow_tables do
 159     Meta.allow_tag_with_these_attributes("table", [])
 160     Meta.allow_tag_with_these_attributes("tbody", [])
 161     Meta.allow_tag_with_these_attributes("td", [])
 162     Meta.allow_tag_with_these_attributes("th", [])
 163     Meta.allow_tag_with_these_attributes("thead", [])
 164     Meta.allow_tag_with_these_attributes("tr", [])
 165   end
 166
 167   @allow_headings Keyword.get(@markup, :allow_headings)
 168
 169   if @allow_headings do
 170     Meta.allow_tag_with_these_attributes("h1", [])
 171     Meta.allow_tag_with_these_attributes("h2", [])
 172     Meta.allow_tag_with_these_attributes("h3", [])
 173     Meta.allow_tag_with_these_attributes("h4", [])
 174     Meta.allow_tag_with_these_attributes("h5", [])
 175   end
 176
 177   @allow_fonts Keyword.get(@markup, :allow_fonts)
 178
 179   if @allow_fonts do
 180     Meta.allow_tag_with_these_attributes("font", ["face"])
 181   end
 182
 183   Meta.strip_everything_not_covered()
 184 end
 185
 186 defmodule Pleroma.HTML.Transform.MediaProxy do
 187   @moduledoc "Transforms inline image URIs to use MediaProxy."
 188
 189   alias Pleroma.Web.MediaProxy
 190
 191   def before_scrub(html), do: html
 192
 193   def scrub_attribute("img", {"src", "http" <> target}) do
 194     media_url =
 195       ("http" <> target)
 196       |> MediaProxy.url()
 197
 198     {"src", media_url}
 199   end
 200
 201   def scrub_attribute(_tag, attribute), do: attribute
 202
 203   def scrub({"img", attributes, children}) do
 204     attributes =
 205       attributes
 206       |> Enum.map(fn attr -> scrub_attribute("img", attr) end)
 207       |> Enum.reject(&is_nil(&1))
 208
 209     {"img", attributes, children}
 210   end
 211
 212   def scrub({:comment, _children}), do: ""
 213
 214   def scrub({tag, attributes, children}), do: {tag, attributes, children}
 215   def scrub({_tag, children}), do: children
 216   def scrub(text), do: text
 217 end