diff --git a/lib/pleroma/html.ex b/lib/pleroma/html.ex
index a0473676b..169394af9 100644
--- a/lib/pleroma/html.ex
+++ b/lib/pleroma/html.ex
@@ -5,18 +5,8 @@
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]
-
- def get_scrubbers() do
- Pleroma.Config.get([:markup, :scrub_policy])
- |> get_scrubbers
- end
-
- def filter_tags(html, nil) do
- get_scrubbers()
- |> Enum.reduce(html, fn scrubber, html ->
+ def filter_tags(html, scrubbers) when is_list(scrubbers) do
+ Enum.reduce(scrubbers, html, fn scrubber, html ->
filter_tags(html, scrubber)
end)
end
@@ -39,6 +29,10 @@ defmodule Pleroma.HTML.Scrubber.TwitterText do
require HtmlSanitizeEx.Scrubber.Meta
alias HtmlSanitizeEx.Scrubber.Meta
+ def version do
+ 0
+ end
+
Meta.remove_cdata_sections_before_scrub()
Meta.strip_comments()
@@ -77,6 +71,10 @@ defmodule Pleroma.HTML.Scrubber.Default do
require HtmlSanitizeEx.Scrubber.Meta
alias HtmlSanitizeEx.Scrubber.Meta
+ def version do
+ 0
+ end
+
@markup Application.get_env(:pleroma, :markup)
@uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
@valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
@@ -152,6 +150,10 @@ end
defmodule Pleroma.HTML.Transform.MediaProxy do
@moduledoc "Transforms inline image URIs to use MediaProxy."
+ def version do
+ 0
+ end
+
alias Pleroma.Web.MediaProxy
def before_scrub(html), do: html
diff --git a/lib/pleroma/user.ex b/lib/pleroma/user.ex
index 1f930479d..b0b65cbe2 100644
--- a/lib/pleroma/user.ex
+++ b/lib/pleroma/user.ex
@@ -733,7 +733,9 @@ defmodule Pleroma.User do
Pleroma.HTML.Scrubber.TwitterText
end
- def html_filter_policy(_), do: nil
+ @default_scrubbers Pleroma.Config.get([:markup, :scrub_policy])
+
+ def html_filter_policy(_), do: @default_scrubbers
def get_or_fetch_by_ap_id(ap_id) do
user = get_by_ap_id(ap_id)
diff --git a/lib/pleroma/web/common_api/common_api.ex b/lib/pleroma/web/common_api/common_api.ex
index 5e5821561..06d44451e 100644
--- a/lib/pleroma/web/common_api/common_api.ex
+++ b/lib/pleroma/web/common_api/common_api.ex
@@ -128,7 +128,8 @@ defmodule Pleroma.Web.CommonAPI do
|> Enum.reduce(%{}, fn {name, file}, acc ->
Map.put(acc, name, "#{Pleroma.Web.Endpoint.static_url()}#{file}")
end)
- ) do
+ ),
+ object <- Map.put(object, "scrubber_cache", %{}) do
res =
ActivityPub.create(%{
to: to,
diff --git a/lib/pleroma/web/common_api/utils.ex b/lib/pleroma/web/common_api/utils.ex
index b91cfc4bb..5c37fd671 100644
--- a/lib/pleroma/web/common_api/utils.ex
+++ b/lib/pleroma/web/common_api/utils.ex
@@ -5,7 +5,7 @@
defmodule Pleroma.Web.CommonAPI.Utils do
alias Calendar.Strftime
alias Comeonin.Pbkdf2
- alias Pleroma.{Activity, Formatter, Object, Repo}
+ alias Pleroma.{Activity, Formatter, Object, Repo, HTML}
alias Pleroma.User
alias Pleroma.Web
alias Pleroma.Web.ActivityPub.Utils
@@ -261,4 +261,63 @@ defmodule Pleroma.Web.CommonAPI.Utils do
}
end)
end
+
+ @doc """
+ Get sanitized HTML from cache, or scrub it and save to cache.
+ """
+ def get_scrubbed_html(
+ content,
+ scrubbers,
+ %{data: %{"object" => object}} = activity
+ ) do
+ scrubber_cache =
+ if object["scrubber_cache"] != nil and is_list(object["scrubber_cache"]) do
+ object["scrubber_cache"]
+ else
+ []
+ end
+
+ key = generate_scrubber_key(scrubbers)
+
+ {new_scrubber_cache, scrubbed_html} =
+ Enum.map_reduce(scrubber_cache, nil, fn %{
+ :scrubbers => current_key,
+ :content => current_content
+ },
+ _ ->
+ if Map.keys(current_key) == Map.keys(key) do
+ if scrubbers == key do
+ {current_key, current_content}
+ else
+ # Remove the entry if scrubber version is outdated
+ {nil, nil}
+ end
+ end
+ end)
+
+ new_scrubber_cache = Enum.reject(new_scrubber_cache, &is_nil/1)
+
+ if !(new_scrubber_cache == scrubber_cache) or scrubbed_html == nil do
+ scrubbed_html = HTML.filter_tags(content, scrubbers)
+ new_scrubber_cache = [%{:scrubbers => key, :content => scrubbed_html} | new_scrubber_cache]
+
+ activity =
+ Map.merge(activity, %{
+ data: %{"object" => %{"scrubber_cache" => new_scrubber_cache}}
+ })
+
+ cng = Ecto.Changeset.change(activity)
+ Repo.update(cng)
+ scrubbed_html
+ else
+ IO.puts("got the post from cache")
+ scrubbed_html
+ end
+ end
+
+ defp generate_scrubber_key(scrubbers) do
+ Enum.reduce(scrubbers, %{}, fn scrubber, acc ->
+ Map.put(acc, to_string(scrubber), scrubber.version)
+ end)
+ end
end
diff --git a/lib/pleroma/web/mastodon_api/views/status_view.ex b/lib/pleroma/web/mastodon_api/views/status_view.ex
index 4d4681da8..8fa3798a6 100644
--- a/lib/pleroma/web/mastodon_api/views/status_view.ex
+++ b/lib/pleroma/web/mastodon_api/views/status_view.ex
@@ -120,7 +120,7 @@ defmodule Pleroma.Web.MastodonAPI.StatusView do
content =
object
|> render_content()
- |> HTML.filter_tags(User.html_filter_policy(opts[:for]))
+ |> Utils.get_scrubbed_html(User.html_filter_policy(opts[:for]), activity)
%{
id: to_string(activity.id),
diff --git a/lib/pleroma/web/twitter_api/views/activity_view.ex b/lib/pleroma/web/twitter_api/views/activity_view.ex
index 592cf622f..adac1dfe9 100644
--- a/lib/pleroma/web/twitter_api/views/activity_view.ex
+++ b/lib/pleroma/web/twitter_api/views/activity_view.ex
@@ -245,7 +245,7 @@ defmodule Pleroma.Web.TwitterAPI.ActivityView do
html =
content
- |> HTML.filter_tags(User.html_filter_policy(opts[:for]))
+ |> Utils.get_scrubbed_html(User.html_filter_policy(opts[:for]), activity)
|> Formatter.emojify(object["emoji"])
text =