1 # Pleroma: A lightweight social networking server
2 # Copyright © 2017-2019 Pleroma Authors <https://pleroma.social/>
3 # SPDX-License-Identifier: AGPL-3.0-only
5 defmodule Pleroma.HTML do
6 alias HtmlSanitizeEx.Scrubber
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]
13 Pleroma.Config.get([:markup, :scrub_policy])
17 def filter_tags(html, nil) do
18 filter_tags(html, get_scrubbers())
21 def filter_tags(html, scrubbers) when is_list(scrubbers) do
22 Enum.reduce(scrubbers, html, fn scrubber, html ->
23 filter_tags(html, scrubber)
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)
31 # TODO: rename object to activity because that's what it is really working with
32 def get_cached_scrubbed_html_for_object(content, scrubbers, object, module) do
33 key = "#{module}#{generate_scrubber_signature(scrubbers)}|#{object.id}"
35 Cachex.fetch!(:scrubber_cache, key, fn _key ->
36 ensure_scrubbed_html(content, scrubbers, object.data["object"]["fake"] || false)
40 def get_cached_stripped_html_for_object(content, object, module) do
41 get_cached_scrubbed_html_for_object(
43 HtmlSanitizeEx.Scrubber.StripTags,
49 def ensure_scrubbed_html(
54 {:commit, filter_tags(content, scrubbers)}
57 def ensure_scrubbed_html(
62 {:ignore, filter_tags(content, scrubbers)}
65 defp generate_scrubber_signature(scrubber) when is_atom(scrubber) do
66 generate_scrubber_signature([scrubber])
69 defp generate_scrubber_signature(scrubbers) do
70 Enum.reduce(scrubbers, "", fn scrubber, signature ->
71 "#{signature}#{to_string(scrubber)}"
75 def extract_first_external_url(_, nil), do: {:error, "No content"}
77 def extract_first_external_url(object, content) do
78 key = "URL|#{object.id}"
80 Cachex.fetch!(:scrubber_cache, key, fn _key ->
83 |> Floki.filter_out("a.mention")
84 |> Floki.attribute("a", "href")
87 {:commit, {:ok, result}}
92 defmodule Pleroma.HTML.Scrubber.TwitterText do
94 An HTML scrubbing policy which limits to twitter-style text. Only
95 paragraphs, breaks and links are allowed through the filter.
98 @markup Application.get_env(:pleroma, :markup)
99 @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
101 require HtmlSanitizeEx.Scrubber.Meta
102 alias HtmlSanitizeEx.Scrubber.Meta
104 Meta.remove_cdata_sections_before_scrub()
105 Meta.strip_comments()
108 Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
109 Meta.allow_tag_with_these_attributes("a", ["name", "title", "class"])
111 Meta.allow_tag_with_this_attribute_values("a", "rel", [
118 # paragraphs and linebreaks
119 Meta.allow_tag_with_these_attributes("br", [])
120 Meta.allow_tag_with_these_attributes("p", [])
123 Meta.allow_tag_with_these_attributes("span", ["class"])
125 # allow inline images for custom emoji
126 @allow_inline_images Keyword.get(@markup, :allow_inline_images)
128 if @allow_inline_images do
129 # restrict img tags to http/https only, because of MediaProxy.
130 Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
132 Meta.allow_tag_with_these_attributes("img", [
140 Meta.strip_everything_not_covered()
143 defmodule Pleroma.HTML.Scrubber.Default do
144 @doc "The default HTML scrubbing policy: no "
146 require HtmlSanitizeEx.Scrubber.Meta
147 alias HtmlSanitizeEx.Scrubber.Meta
148 # credo:disable-for-previous-line
149 # No idea how to fix this one…
151 @markup Application.get_env(:pleroma, :markup)
152 @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
154 Meta.remove_cdata_sections_before_scrub()
155 Meta.strip_comments()
157 Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
158 Meta.allow_tag_with_these_attributes("a", ["name", "title", "class"])
160 Meta.allow_tag_with_this_attribute_values("a", "rel", [
167 Meta.allow_tag_with_these_attributes("abbr", ["title"])
169 Meta.allow_tag_with_these_attributes("b", [])
170 Meta.allow_tag_with_these_attributes("blockquote", [])
171 Meta.allow_tag_with_these_attributes("br", [])
172 Meta.allow_tag_with_these_attributes("code", [])
173 Meta.allow_tag_with_these_attributes("del", [])
174 Meta.allow_tag_with_these_attributes("em", [])
175 Meta.allow_tag_with_these_attributes("i", [])
176 Meta.allow_tag_with_these_attributes("li", [])
177 Meta.allow_tag_with_these_attributes("ol", [])
178 Meta.allow_tag_with_these_attributes("p", [])
179 Meta.allow_tag_with_these_attributes("pre", [])
180 Meta.allow_tag_with_these_attributes("span", ["class"])
181 Meta.allow_tag_with_these_attributes("strong", [])
182 Meta.allow_tag_with_these_attributes("u", [])
183 Meta.allow_tag_with_these_attributes("ul", [])
185 @allow_inline_images Keyword.get(@markup, :allow_inline_images)
187 if @allow_inline_images do
188 # restrict img tags to http/https only, because of MediaProxy.
189 Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
191 Meta.allow_tag_with_these_attributes("img", [
199 @allow_tables Keyword.get(@markup, :allow_tables)
202 Meta.allow_tag_with_these_attributes("table", [])
203 Meta.allow_tag_with_these_attributes("tbody", [])
204 Meta.allow_tag_with_these_attributes("td", [])
205 Meta.allow_tag_with_these_attributes("th", [])
206 Meta.allow_tag_with_these_attributes("thead", [])
207 Meta.allow_tag_with_these_attributes("tr", [])
210 @allow_headings Keyword.get(@markup, :allow_headings)
212 if @allow_headings do
213 Meta.allow_tag_with_these_attributes("h1", [])
214 Meta.allow_tag_with_these_attributes("h2", [])
215 Meta.allow_tag_with_these_attributes("h3", [])
216 Meta.allow_tag_with_these_attributes("h4", [])
217 Meta.allow_tag_with_these_attributes("h5", [])
220 @allow_fonts Keyword.get(@markup, :allow_fonts)
223 Meta.allow_tag_with_these_attributes("font", ["face"])
226 Meta.strip_everything_not_covered()
229 defmodule Pleroma.HTML.Transform.MediaProxy do
230 @moduledoc "Transforms inline image URIs to use MediaProxy."
232 alias Pleroma.Web.MediaProxy
234 def before_scrub(html), do: html
236 def scrub_attribute("img", {"src", "http" <> target}) do
244 def scrub_attribute(_tag, attribute), do: attribute
246 def scrub({"img", attributes, children}) do
249 |> Enum.map(fn attr -> scrub_attribute("img", attr) end)
250 |> Enum.reject(&is_nil(&1))
252 {"img", attributes, children}
255 def scrub({:comment, _children}), do: ""
257 def scrub({tag, attributes, children}), do: {tag, attributes, children}
258 def scrub({_tag, children}), do: children
259 def scrub(text), do: text