X-Git-Url: http://git.squeep.com/?a=blobdiff_plain;f=lib%2Fpleroma%2Fhtml.ex;h=2cae29f35170fe5b952551b405db677c8baba9f2;hb=92b4a1aa1bc750bb077ae45c422967f9712e247d;hp=e5e78ee4f50124a1d1c2d0568b5d7c14f6c0fa67;hpb=ad5263c647aea65dbeb4c329825671895e0a8863;p=akkoma

diff --git a/lib/pleroma/html.ex b/lib/pleroma/html.ex
index e5e78ee4f..2cae29f35 100644
--- a/lib/pleroma/html.ex
+++ b/lib/pleroma/html.ex
@@ -3,7 +3,24 @@
 # SPDX-License-Identifier: AGPL-3.0-only
 
 defmodule Pleroma.HTML do
-  alias HtmlSanitizeEx.Scrubber
+  # Scrubbers are compiled on boot so they can be configured in OTP releases
+  #  @on_load :compile_scrubbers
+
+  def compile_scrubbers do
+    dir = Path.join(:code.priv_dir(:pleroma), "scrubbers")
+
+    dir
+    |> File.ls!()
+    |> Enum.map(&Path.join(dir, &1))
+    |> Kernel.ParallelCompiler.compile()
+    |> case do
+      {:error, _errors, _warnings} ->
+        raise "Compiling scrubbers failed"
+
+      {:ok, _modules, _warnings} ->
+        :ok
+    end
+  end
 
   defp get_scrubbers(scrubber) when is_atom(scrubber), do: [scrubber]
   defp get_scrubbers(scrubbers) when is_list(scrubbers), do: scrubbers
@@ -24,9 +41,13 @@ defmodule Pleroma.HTML do
     end)
   end
 
-  def filter_tags(html, scrubber), do: Scrubber.scrub(html, scrubber)
+  def filter_tags(html, scrubber) do
+    {:ok, content} = FastSanitize.Sanitizer.scrub(html, scrubber)
+    content
+  end
+
   def filter_tags(html), do: filter_tags(html, nil)
-  def strip_tags(html), do: Scrubber.scrub(html, Scrubber.StripTags)
+  def strip_tags(html), do: filter_tags(html, FastSanitize.Sanitizer.StripTags)
 
   def get_cached_scrubbed_html_for_activity(
         content,
@@ -46,7 +67,7 @@ defmodule Pleroma.HTML do
   def get_cached_stripped_html_for_activity(content, activity, key) do
     get_cached_scrubbed_html_for_activity(
       content,
-      HtmlSanitizeEx.Scrubber.StripTags,
+      FastSanitize.Sanitizer.StripTags,
       activity,
       key,
       &HtmlEntities.decode/1
@@ -89,7 +110,7 @@ defmodule Pleroma.HTML do
     Cachex.fetch!(:scrubber_cache, key, fn _key ->
       result =
         content
-        |> Floki.filter_out("a.mention")
+        |> Floki.filter_out("a.mention,a.hashtag,a[rel~=\"tag\"]")
         |> Floki.attribute("a", "href")
         |> Enum.at(0)
 
@@ -97,186 +118,3 @@ defmodule Pleroma.HTML do
     end)
   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.
-  """
-
-  @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
-
-  require HtmlSanitizeEx.Scrubber.Meta
-  alias HtmlSanitizeEx.Scrubber.Meta
-
-  Meta.remove_cdata_sections_before_scrub()
-  Meta.strip_comments()
-
-  # links
-  Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
-
-  Meta.allow_tag_with_this_attribute_values("a", "class", [
-    "hashtag",
-    "u-url",
-    "mention",
-    "u-url mention",
-    "mention u-url"
-  ])
-
-  Meta.allow_tag_with_this_attribute_values("a", "rel", [
-    "tag",
-    "nofollow",
-    "noopener",
-    "noreferrer"
-  ])
-
-  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_this_attribute_values("span", "class", ["h-card"])
-  Meta.allow_tag_with_these_attributes("span", [])
-
-  # allow inline images for custom emoji
-  if Pleroma.Config.get([:markup, :allow_inline_images]) do
-    # 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",
-      "height",
-      "class",
-      "title",
-      "alt"
-    ])
-  end
-
-  Meta.strip_everything_not_covered()
-end
-
-defmodule Pleroma.HTML.Scrubber.Default do
-  @doc "The default HTML scrubbing policy: no "
-
-  require HtmlSanitizeEx.Scrubber.Meta
-  alias HtmlSanitizeEx.Scrubber.Meta
-  # credo:disable-for-previous-line
-  # No idea how to fix this one…
-
-  @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
-
-  Meta.remove_cdata_sections_before_scrub()
-  Meta.strip_comments()
-
-  Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
-
-  Meta.allow_tag_with_this_attribute_values("a", "class", [
-    "hashtag",
-    "u-url",
-    "mention",
-    "u-url mention",
-    "mention u-url"
-  ])
-
-  Meta.allow_tag_with_this_attribute_values("a", "rel", [
-    "tag",
-    "nofollow",
-    "noopener",
-    "noreferrer"
-  ])
-
-  Meta.allow_tag_with_these_attributes("a", ["name", "title"])
-
-  Meta.allow_tag_with_these_attributes("abbr", ["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("strong", [])
-  Meta.allow_tag_with_these_attributes("u", [])
-  Meta.allow_tag_with_these_attributes("ul", [])
-
-  Meta.allow_tag_with_this_attribute_values("span", "class", ["h-card"])
-  Meta.allow_tag_with_these_attributes("span", [])
-
-  @allow_inline_images Pleroma.Config.get([:markup, :allow_inline_images])
-
-  if @allow_inline_images do
-    # 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",
-      "height",
-      "class",
-      "title",
-      "alt"
-    ])
-  end
-
-  if Pleroma.Config.get([:markup, :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
-
-  if Pleroma.Config.get([:markup, :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
-
-  if Pleroma.Config.get([:markup, :allow_fonts]) do
-    Meta.allow_tag_with_these_attributes("font", ["face"])
-  end
-
-  Meta.strip_everything_not_covered()
-end
-
-defmodule Pleroma.HTML.Transform.MediaProxy do
-  @moduledoc "Transforms inline image URIs to use MediaProxy."
-
-  alias Pleroma.Web.MediaProxy
-
-  def before_scrub(html), do: html
-
-  def scrub_attribute("img", {"src", "http" <> target}) do
-    media_url =
-      ("http" <> target)
-      |> MediaProxy.url()
-
-    {"src", media_url}
-  end
-
-  def scrub_attribute(_tag, attribute), do: attribute
-
-  def scrub({"img", attributes, children}) do
-    attributes =
-      attributes
-      |> Enum.map(fn attr -> scrub_attribute("img", attr) end)
-      |> Enum.reject(&is_nil(&1))
-
-    {"img", attributes, children}
-  end
-
-  def scrub({:comment, _children}), do: ""
-
-  def scrub({tag, attributes, children}), do: {tag, attributes, children}
-  def scrub({_tag, children}), do: children
-  def scrub(text), do: text
-end