From 08f68370659597d6bc428e425925bcb9516d5706 Mon Sep 17 00:00:00 2001 From: rinpatch Date: Tue, 29 Oct 2019 01:18:08 +0300 Subject: [PATCH] Switch from HtmlSanitizeEx to FastSanitize --- lib/pleroma/html.ex | 129 +++++++++--------- test/emoji/formatter_test.exs | 2 +- test/html_test.exs | 44 +++--- .../mrf/normalize_markup_test.exs | 10 +- test/web/common_api/common_api_test.exs | 4 +- 5 files changed, 95 insertions(+), 94 deletions(-) diff --git a/lib/pleroma/html.ex b/lib/pleroma/html.ex index 937bafed5..fd0495049 100644 --- a/lib/pleroma/html.ex +++ b/lib/pleroma/html.ex @@ -3,7 +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 @@ -24,9 +23,13 @@ def filter_tags(html, scrubbers) when is_list(scrubbers) 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, @@ -36,7 +39,6 @@ def get_cached_scrubbed_html_for_activity( callback \\ fn x -> x end ) do key = "#{key}#{generate_scrubber_signature(scrubbers)}|#{activity.id}" - Cachex.fetch!(:scrubber_cache, key, fn _key -> object = Pleroma.Object.normalize(activity) ensure_scrubbed_html(content, scrubbers, object.data["fake"] || false, callback) @@ -46,7 +48,7 @@ def get_cached_scrubbed_html_for_activity( 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 @@ -109,13 +111,12 @@ defmodule Pleroma.HTML.Scrubber.TwitterText do 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_uri_attributes(:a, ["href", "data-user", "data-tag"], @valid_schemes) - Meta.allow_tag_with_this_attribute_values("a", "class", [ + Meta.allow_tag_with_this_attribute_values(:a, "class", [ "hashtag", "u-url", "mention", @@ -123,29 +124,29 @@ defmodule Pleroma.HTML.Scrubber.TwitterText do "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"]) + 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_this_attribute_values("span", "class", ["h-card"]) - Meta.allow_tag_with_these_attributes("span", []) + 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_uri_attributes(:img, ["src"], ["http", "https"]) - Meta.allow_tag_with_these_attributes("img", [ + Meta.allow_tag_with_these_attributes(:img, [ "width", "height", "class", @@ -160,19 +161,19 @@ defmodule Pleroma.HTML.Scrubber.TwitterText do 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… @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], []) - Meta.remove_cdata_sections_before_scrub() +# 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_uri_attributes(:a, ["href", "data-user", "data-tag"], @valid_schemes) - Meta.allow_tag_with_this_attribute_values("a", "class", [ + Meta.allow_tag_with_this_attribute_values(:a, "class", [ "hashtag", "u-url", "mention", @@ -180,7 +181,7 @@ defmodule Pleroma.HTML.Scrubber.Default do "mention u-url" ]) - Meta.allow_tag_with_this_attribute_values("a", "rel", [ + Meta.allow_tag_with_this_attribute_values(:a, "rel", [ "tag", "nofollow", "noopener", @@ -188,37 +189,37 @@ defmodule Pleroma.HTML.Scrubber.Default do "ugc" ]) - Meta.allow_tag_with_these_attributes("a", ["name", "title"]) + Meta.allow_tag_with_these_attributes(:a, ["name", "title"]) - Meta.allow_tag_with_these_attributes("abbr", ["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_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", []) + 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", @@ -228,24 +229,24 @@ defmodule Pleroma.HTML.Scrubber.Default do 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", []) + 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", []) + 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"]) + Meta.allow_tag_with_these_attributes(:font, ["face"]) end Meta.strip_everything_not_covered() @@ -258,7 +259,7 @@ 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() @@ -268,16 +269,16 @@ def scrub_attribute("img", {"src", "http" <> target}) 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 @@ -298,9 +299,9 @@ defmodule Pleroma.HTML.Scrubber.LinksOnly do Meta.strip_comments() # links - Meta.allow_tag_with_uri_attributes("a", ["href"], @valid_schemes) + Meta.allow_tag_with_uri_attributes(:a, ["href"], @valid_schemes) - Meta.allow_tag_with_this_attribute_values("a", "rel", [ + Meta.allow_tag_with_this_attribute_values(:a, "rel", [ "tag", "nofollow", "noopener", @@ -309,6 +310,6 @@ defmodule Pleroma.HTML.Scrubber.LinksOnly do "ugc" ]) - Meta.allow_tag_with_these_attributes("a", ["name", "title"]) + Meta.allow_tag_with_these_attributes(:a, ["name", "title"]) Meta.strip_everything_not_covered() end diff --git a/test/emoji/formatter_test.exs b/test/emoji/formatter_test.exs index 6d25fc453..3e37ae3f0 100644 --- a/test/emoji/formatter_test.exs +++ b/test/emoji/formatter_test.exs @@ -12,7 +12,7 @@ test "it adds cool emoji" do text = "I love :firefox:" expected_result = - "I love \"firefox\"" + "I love \"firefox\"" assert Formatter.emojify(text) == expected_result end diff --git a/test/html_test.exs b/test/html_test.exs index 306ad3b3b..f0869534c 100644 --- a/test/html_test.exs +++ b/test/html_test.exs @@ -21,31 +21,31 @@ defmodule Pleroma.HTMLTest do """ @html_onerror_sample """ - + """ @html_span_class_sample """ - hi + hi """ @html_span_microformats_sample """ - @foo + @foo """ @html_span_invalid_microformats_sample """ - @foo + @foo """ describe "StripTags scrubber" do test "works as expected" do expected = """ - this is in bold + this is in bold this is a paragraph this is a linebreak - this is a link with allowed "rel" attribute: example.com - this is a link with not allowed "rel" attribute: example.com + this is a link with allowed "rel" attribute: example.com + this is a link with not allowed "rel" attribute: example.com this is an image: - alert('hacked') + alert('hacked') """ assert expected == HTML.strip_tags(@html_sample) @@ -61,13 +61,13 @@ test "does not allow attribute-based XSS" do describe "TwitterText scrubber" do test "normalizes HTML as expected" do expected = """ - this is in bold + this is in bold

this is a paragraph

- this is a linebreak
- this is a link with allowed "rel" attribute: - this is a link with not allowed "rel" attribute: example.com - this is an image:
- alert('hacked') + this is a linebreak
+ this is a link with allowed "rel" attribute: + this is a link with not allowed "rel" attribute: example.com + this is an image:
+ alert('hacked') """ assert expected == HTML.filter_tags(@html_sample, Pleroma.HTML.Scrubber.TwitterText) @@ -75,7 +75,7 @@ test "normalizes HTML as expected" do test "does not allow attribute-based XSS" do expected = """ - + """ assert expected == HTML.filter_tags(@html_onerror_sample, Pleroma.HTML.Scrubber.TwitterText) @@ -115,13 +115,13 @@ test "filters invalid microformats markup" do describe "default scrubber" do test "normalizes HTML as expected" do expected = """ - this is in bold + this is in bold

this is a paragraph

- this is a linebreak
- this is a link with allowed "rel" attribute: - this is a link with not allowed "rel" attribute: example.com - this is an image:
- alert('hacked') + this is a linebreak
+ this is a link with allowed "rel" attribute: + this is a link with not allowed "rel" attribute: example.com + this is an image:
+ alert('hacked') """ assert expected == HTML.filter_tags(@html_sample, Pleroma.HTML.Scrubber.Default) @@ -129,7 +129,7 @@ test "normalizes HTML as expected" do test "does not allow attribute-based XSS" do expected = """ - + """ assert expected == HTML.filter_tags(@html_onerror_sample, Pleroma.HTML.Scrubber.Default) diff --git a/test/web/activity_pub/mrf/normalize_markup_test.exs b/test/web/activity_pub/mrf/normalize_markup_test.exs index 3916a1f35..0207be56b 100644 --- a/test/web/activity_pub/mrf/normalize_markup_test.exs +++ b/test/web/activity_pub/mrf/normalize_markup_test.exs @@ -20,11 +20,11 @@ test "it filter html tags" do expected = """ this is in bold

this is a paragraph

- this is a linebreak
- this is a link with allowed "rel" attribute: - this is a link with not allowed "rel" attribute: example.com - this is an image:
- alert('hacked') + this is a linebreak
+ this is a link with allowed "rel" attribute: + this is a link with not allowed "rel" attribute: example.com + this is an image:
+ alert('hacked') """ message = %{"type" => "Create", "object" => %{"content" => @html_sample}} diff --git a/test/web/common_api/common_api_test.exs b/test/web/common_api/common_api_test.exs index 1d2f20617..212b00cbb 100644 --- a/test/web/common_api/common_api_test.exs +++ b/test/web/common_api/common_api_test.exs @@ -140,7 +140,7 @@ test "it filters out obviously bad tags when accepting a post as HTML" do object = Object.normalize(activity) - assert object.data["content"] == "

2hu

alert('xss')" + assert object.data["content"] == "

2hu

alert('xss')" end test "it filters out obviously bad tags when accepting a post as Markdown" do @@ -156,7 +156,7 @@ test "it filters out obviously bad tags when accepting a post as Markdown" do object = Object.normalize(activity) - assert object.data["content"] == "

2hu

alert('xss')" + assert object.data["content"] == "

2hu

alert('xss')" end test "it does not allow replies to direct messages that are not direct messages themselves" do