X-Git-Url: https://git.squeep.com/?a=blobdiff_plain;f=lib%2Fpleroma%2Fhtml.ex;h=d78c5f2025ced17c45649393954cc164841dd8d2;hb=93f64aba78f130cced2fa30dd3475e5c446be9d6;hp=7f1dbe28c81ce1e9a94cb43ae96bc0291d8cfb17;hpb=bffddf5e31c53d0603de387c61b6011f70785b97;p=akkoma
diff --git a/lib/pleroma/html.ex b/lib/pleroma/html.ex
index 7f1dbe28c..d78c5f202 100644
--- a/lib/pleroma/html.ex
+++ b/lib/pleroma/html.ex
@@ -1,9 +1,24 @@
# Pleroma: A lightweight social networking server
-# Copyright © 2017-2019 Pleroma Authors
+# Copyright © 2017-2020 Pleroma Authors
# 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
+ |> Pleroma.Utils.compile_dir()
+ |> 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
@@ -24,41 +39,55 @@ 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
+ fake,
+ callback
) do
- {:commit, filter_tags(content, scrubbers)}
- end
+ content =
+ content
+ |> filter_tags(scrubbers)
+ |> callback.()
- def ensure_scrubbed_html(
- content,
- scrubbers,
- true = _fake
- ) do
- {:ignore, filter_tags(content, scrubbers)}
+ if fake do
+ {:ignore, content}
+ else
+ {:commit, content}
+ end
end
defp generate_scrubber_signature(scrubber) when is_atom(scrubber) do
@@ -79,7 +108,8 @@ defmodule Pleroma.HTML do
Cachex.fetch!(:scrubber_cache, key, fn _key ->
result =
content
- |> Floki.filter_out("a.mention")
+ |> Floki.parse_fragment!()
+ |> Floki.filter_out("a.mention,a.hashtag,a[rel~=\"tag\"]")
|> Floki.attribute("a", "href")
|> Enum.at(0)
@@ -87,173 +117,3 @@ 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.
- """
-
- @markup Application.get_env(:pleroma, :markup)
- @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_these_attributes("a", ["name", "title", "class"])
-
- Meta.allow_tag_with_this_attribute_values("a", "rel", [
- "tag",
- "nofollow",
- "noopener",
- "noreferrer"
- ])
-
- # 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", ["class"])
-
- # allow inline images for custom emoji
- @allow_inline_images Keyword.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",
- "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 oneâ¦
-
- @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_this_attribute_values("a", "rel", [
- "tag",
- "nofollow",
- "noopener",
- "noreferrer"
- ])
-
- 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)
-
- 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",
- "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
-
-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