+# Pleroma: A lightweight social networking server
+# Copyright © 2017-2019 Pleroma Authors <https://pleroma.social/>
+# SPDX-License-Identifier: AGPL-3.0-only
+
defmodule Pleroma.HTML do
alias HtmlSanitizeEx.Scrubber
- @markup Application.get_env(:pleroma, :markup)
+ 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
+ filter_tags(html, get_scrubbers())
+ end
+
+ def filter_tags(html, scrubbers) when is_list(scrubbers) do
+ Enum.reduce(scrubbers, html, fn scrubber, html ->
+ filter_tags(html, scrubber)
+ end)
+ end
+
+ def filter_tags(html, scrubber), do: Scrubber.scrub(html, scrubber)
+ def filter_tags(html), do: filter_tags(html, nil)
+ def strip_tags(html), do: Scrubber.scrub(html, Scrubber.StripTags)
+
+ def get_cached_scrubbed_html_for_object(content, scrubbers, object) do
+ key = "#{generate_scrubber_signature(scrubbers)}|#{object.id}"
+ Cachex.fetch!(:scrubber_cache, key, fn _key -> ensure_scrubbed_html(content, scrubbers) end)
+ end
- def filter_tags(html) do
- scrubber = Keyword.get(@markup, :scrub_policy)
- html |> Scrubber.scrub(scrubber)
+ def get_cached_stripped_html_for_object(content, object) do
+ get_cached_scrubbed_html_for_object(content, HtmlSanitizeEx.Scrubber.StripTags, object)
end
- def strip_tags(html) do
- html |> Scrubber.scrub(Scrubber.StripTags)
+ def ensure_scrubbed_html(
+ content,
+ scrubbers
+ ) do
+ {:commit, filter_tags(content, scrubbers)}
+ end
+
+ defp generate_scrubber_signature(scrubber) when is_atom(scrubber) do
+ generate_scrubber_signature([scrubber])
+ end
+
+ defp generate_scrubber_signature(scrubbers) do
+ Enum.reduce(scrubbers, "", fn scrubber, signature ->
+ # If a scrubber does not have a version(e.g HtmlSanitizeEx.Scrubber.StripTags) it is assumed it is always 0)
+ version =
+ if Kernel.function_exported?(scrubber, :version, 0) do
+ scrubber.version
+ else
+ 0
+ end
+
+ "#{signature}#{to_string(scrubber)}#{version}"
+ end)
end
end
paragraphs, breaks and links are allowed through the filter.
"""
+ @markup Application.get_env(:pleroma, :markup)
+ @uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
+ @valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
+
require HtmlSanitizeEx.Scrubber.Meta
alias HtmlSanitizeEx.Scrubber.Meta
- @valid_schemes ["http", "https"]
+ def version do
+ 0
+ end
Meta.remove_cdata_sections_before_scrub()
Meta.strip_comments()
# links
- Meta.allow_tag_with_uri_attributes("a", ["href"], @valid_schemes)
+ Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
Meta.allow_tag_with_these_attributes("a", ["name", "title"])
# paragraphs and linebreaks
Meta.allow_tag_with_these_attributes("span", [])
# allow inline images for custom emoji
- @markup Application.get_env(:pleroma, :markup)
@allow_inline_images Keyword.get(@markup, :allow_inline_images)
if @allow_inline_images do
- Meta.allow_tag_with_uri_attributes("img", ["src"], @valid_schemes)
+ # 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",
"alt"
])
end
+
+ Meta.strip_everything_not_covered()
end
defmodule Pleroma.HTML.Scrubber.Default do
require HtmlSanitizeEx.Scrubber.Meta
alias HtmlSanitizeEx.Scrubber.Meta
- @valid_schemes ["http", "https"]
+ 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, [])
Meta.remove_cdata_sections_before_scrub()
Meta.strip_comments()
- Meta.allow_tag_with_uri_attributes("a", ["href"], @valid_schemes)
+ Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
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("u", [])
Meta.allow_tag_with_these_attributes("ul", [])
- @markup Application.get_env(:pleroma, :markup)
@allow_inline_images Keyword.get(@markup, :allow_inline_images)
if @allow_inline_images do
- Meta.allow_tag_with_uri_attributes("img", ["src"], @valid_schemes)
+ # 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",
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 version do
+ 0
+ end
+
+ 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