X-Git-Url: http://git.squeep.com/?a=blobdiff_plain;f=lib%2Fpleroma%2Fhtml.ex;h=997e965f0829afdf4e475a1ecdc16e5f8a9e80b9;hb=19d228cc586a1304ef6e982a447a77f8c3a48772;hp=7f1dbe28c81ce1e9a94cb43ae96bc0291d8cfb17;hpb=0f2f7d2cec8297b1b5645643d7584cde561ce628;p=akkoma

diff --git a/lib/pleroma/html.ex b/lib/pleroma/html.ex
index 7f1dbe28c..997e965f0 100644
--- a/lib/pleroma/html.ex
+++ b/lib/pleroma/html.ex
@@ -3,8 +3,6 @@
 # SPDX-License-Identifier: AGPL-3.0-only
 
 defmodule Pleroma.HTML do
-  alias HtmlSanitizeEx.Scrubber
-
   defp get_scrubbers(scrubber) when is_atom(scrubber), do: [scrubber]
   defp get_scrubbers(scrubbers) when is_list(scrubbers), do: scrubbers
   defp get_scrubbers(_), do: [Pleroma.HTML.Scrubber.Default]
@@ -24,41 +22,55 @@ 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, scrubbers, activity, key \\ "") do
+  def get_cached_scrubbed_html_for_activity(
+        content,
+        scrubbers,
+        activity,
+        key \\ "",
+        callback \\ fn x -> x end
+      ) do
     key = "#{key}#{generate_scrubber_signature(scrubbers)}|#{activity.id}"
 
     Cachex.fetch!(:scrubber_cache, key, fn _key ->
-      ensure_scrubbed_html(content, scrubbers, activity.data["object"]["fake"] || false)
+      object = Pleroma.Object.normalize(activity)
+      ensure_scrubbed_html(content, scrubbers, object.data["fake"] || false, callback)
     end)
   end
 
   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
+      key,
+      &HtmlEntities.decode/1
     )
   end
 
   def ensure_scrubbed_html(
         content,
         scrubbers,
-        false = _fake
-      ) do
-    {:commit, filter_tags(content, scrubbers)}
-  end
-
-  def ensure_scrubbed_html(
-        content,
-        scrubbers,
-        true = _fake
+        fake,
+        callback
       ) do
-    {:ignore, filter_tags(content, scrubbers)}
+    content =
+      content
+      |> filter_tags(scrubbers)
+      |> callback.()
+
+    if fake do
+      {:ignore, content}
+    else
+      {:commit, content}
+    end
   end
 
   defp generate_scrubber_signature(scrubber) when is_atom(scrubber) do
@@ -79,7 +91,7 @@ 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)
 
@@ -94,43 +106,50 @@ defmodule Pleroma.HTML.Scrubber.TwitterText do
   paragraphs, breaks and links are allowed through the filter.
   """
 
-  @markup Application.get_env(:pleroma, :markup)
   @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
 
-  require HtmlSanitizeEx.Scrubber.Meta
-  alias HtmlSanitizeEx.Scrubber.Meta
+  require FastSanitize.Sanitizer.Meta
+  alias FastSanitize.Sanitizer.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_these_attributes("a", ["name", "title", "class"])
+  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", [
+  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", [])
+  Meta.allow_tag_with_these_attributes(:br, [])
+  Meta.allow_tag_with_these_attributes(:p, [])
 
   # microformats
-  Meta.allow_tag_with_these_attributes("span", ["class"])
+  Meta.allow_tag_with_this_attribute_values(:span, "class", ["h-card"])
+  Meta.allow_tag_with_these_attributes(:span, [])
 
   # allow inline images for custom emoji
-  @allow_inline_images Keyword.get(@markup, :allow_inline_images)
-
-  if @allow_inline_images do
+  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_uri_attributes(:img, ["src"], ["http", "https"])
 
-    Meta.allow_tag_with_these_attributes("img", [
+    Meta.allow_tag_with_these_attributes(:img, [
       "width",
       "height",
+      "class",
       "title",
       "alt"
     ])
@@ -142,84 +161,91 @@ end
 defmodule Pleroma.HTML.Scrubber.Default do
   @doc "The default HTML scrubbing policy: no "
 
-  require HtmlSanitizeEx.Scrubber.Meta
-  alias HtmlSanitizeEx.Scrubber.Meta
+  require FastSanitize.Sanitizer.Meta
+  alias FastSanitize.Sanitizer.Meta
   # credo:disable-for-previous-line
   # No idea how to fix this one…
 
-  @markup Application.get_env(:pleroma, :markup)
   @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_these_attributes("a", ["name", "title", "class"])
+  Meta.allow_tag_with_uri_attributes(:a, ["href", "data-user", "data-tag"], @valid_schemes)
 
-  Meta.allow_tag_with_this_attribute_values("a", "rel", [
+  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"
+    "noreferrer",
+    "ugc"
   ])
 
-  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("span", ["class"])
-  Meta.allow_tag_with_these_attributes("strong", [])
-  Meta.allow_tag_with_these_attributes("u", [])
-  Meta.allow_tag_with_these_attributes("ul", [])
-
-  @allow_inline_images Keyword.get(@markup, :allow_inline_images)
+  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(:sub, [])
+  Meta.allow_tag_with_these_attributes(:sup, [])
+  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_uri_attributes(:img, ["src"], ["http", "https"])
 
-    Meta.allow_tag_with_these_attributes("img", [
+    Meta.allow_tag_with_these_attributes(:img, [
       "width",
       "height",
+      "class",
       "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", [])
+  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
 
-  @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", [])
+  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
 
-  @allow_fonts Keyword.get(@markup, :allow_fonts)
-
-  if @allow_fonts do
-    Meta.allow_tag_with_these_attributes("font", ["face"])
+  if Pleroma.Config.get([:markup, :allow_fonts]) do
+    Meta.allow_tag_with_these_attributes(:font, ["face"])
   end
 
   Meta.strip_everything_not_covered()
@@ -232,7 +258,7 @@ defmodule Pleroma.HTML.Transform.MediaProxy do
 
   def before_scrub(html), do: html
 
-  def scrub_attribute("img", {"src", "http" <> target}) do
+  def scrub_attribute(:img, {"src", "http" <> target}) do
     media_url =
       ("http" <> target)
       |> MediaProxy.url()
@@ -242,18 +268,46 @@ defmodule Pleroma.HTML.Transform.MediaProxy do
 
   def scrub_attribute(_tag, attribute), do: attribute
 
-  def scrub({"img", attributes, children}) do
+  def scrub({:img, attributes, children}) do
     attributes =
       attributes
-      |> Enum.map(fn attr -> scrub_attribute("img", attr) end)
+      |> Enum.map(fn attr -> scrub_attribute(:img, attr) end)
       |> Enum.reject(&is_nil(&1))
 
-    {"img", attributes, children}
+    {:img, attributes, children}
   end
 
-  def scrub({:comment, _children}), do: ""
+  def scrub({:comment, _text, _children}), do: ""
 
   def scrub({tag, attributes, children}), do: {tag, attributes, children}
   def scrub({_tag, children}), do: children
   def scrub(text), do: text
 end
+
+defmodule Pleroma.HTML.Scrubber.LinksOnly do
+  @moduledoc """
+  An HTML scrubbing policy which limits to links only.
+  """
+
+  @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
+
+  require FastSanitize.Sanitizer.Meta
+  alias FastSanitize.Sanitizer.Meta
+
+  Meta.strip_comments()
+
+  # links
+  Meta.allow_tag_with_uri_attributes(:a, ["href"], @valid_schemes)
+
+  Meta.allow_tag_with_this_attribute_values(:a, "rel", [
+    "tag",
+    "nofollow",
+    "noopener",
+    "noreferrer",
+    "me",
+    "ugc"
+  ])
+
+  Meta.allow_tag_with_these_attributes(:a, ["name", "title"])
+  Meta.strip_everything_not_covered()
+end