Add advanced text formatting back into glitch-soc

6 files changed, 459 insertions, 29 deletions

diff --git a/app/helpers/formatting_helper.rb b/app/helpers/formatting_helper.rb
index e11156999..2a622ae0b 100644
--- a/app/helpers/formatting_helper.rb
+++ b/app/helpers/formatting_helper.rb
@@ -14,7 +14,7 @@ module FormattingHelper
   end
 
   def status_content_format(status)
-    html_aware_format(status.text, status.local?, preloaded_accounts: [status.account] + (status.respond_to?(:active_mentions) ? status.active_mentions.map(&:account) : []))
+    html_aware_format(status.text, status.local?, preloaded_accounts: [status.account] + (status.respond_to?(:active_mentions) ? status.active_mentions.map(&:account) : []), content_type: status.content_type)
   end
 
   def account_bio_format(account)
diff --git a/app/lib/advanced_text_formatter.rb b/app/lib/advanced_text_formatter.rb
new file mode 100644
index 000000000..5ce87d306
--- /dev/null
+++ b/app/lib/advanced_text_formatter.rb
@@ -0,0 +1,131 @@
+# frozen_string_literal: true
+
+class AdvancedTextFormatter < TextFormatter
+  class HTMLRenderer < Redcarpet::Render::HTML
+    def initialize(options, &block)
+      super(options)
+      @format_link = block
+    end
+
+    def block_code(code, _language)
+      <<~HTML.squish
+        <pre><code>#{h(code).gsub("\n", '<br/>')}</code></pre>
+      HTML
+    end
+
+    def autolink(link, link_type)
+      return link if link_type == :email
+      @format_link.call(link)
+    end
+  end
+
+  # @param [String] text
+  # @param [Hash] options
+  # @option options [Boolean] :multiline
+  # @option options [Boolean] :with_domains
+  # @option options [Boolean] :with_rel_me
+  # @option options [Array<Account>] :preloaded_accounts
+  # @option options [String] :content_type
+  def initialize(text, options = {})
+    content_type = options.delete(:content_type)
+    super(text, options)
+
+    @text = format_markdown(text) if content_type == 'text/markdown'
+  end
+
+  # Differs from TextFormatter by not messing with newline after parsing
+  def to_s
+    return ''.html_safe if text.blank?
+
+    html = rewrite do |entity|
+      if entity[:url]
+        link_to_url(entity)
+      elsif entity[:hashtag]
+        link_to_hashtag(entity)
+      elsif entity[:screen_name]
+        link_to_mention(entity)
+      end
+    end
+
+    html.html_safe # rubocop:disable Rails/OutputSafety
+  end
+
+  # Differs from `TextFormatter` by skipping HTML tags and entities
+  def entities
+    @entities ||= begin
+      gaps = []
+      total_offset = 0
+
+      escaped = text.gsub(/<[^>]*>|&#[0-9]+;/) do |match|
+        total_offset += match.length - 1
+        end_offset = Regexp.last_match.end(0)
+        gaps << [end_offset - total_offset, total_offset]
+        ' '
+      end
+
+      Extractor.extract_entities_with_indices(escaped, extract_url_without_protocol: false).map do |entity|
+        start_pos, end_pos = entity[:indices]
+        offset_idx = gaps.rindex { |gap| gap.first <= start_pos }
+        offset = offset_idx.nil? ? 0 : gaps[offset_idx].last
+        entity.merge(indices: [start_pos + offset, end_pos + offset])
+      end
+    end
+  end
+
+  private
+
+  # Differs from `TextFormatter` in that it keeps HTML; but it sanitizes at the end to remain safe
+  def rewrite
+    entities.sort_by! do |entity|
+      entity[:indices].first
+    end
+
+    result = ''.dup
+
+    last_index = entities.reduce(0) do |index, entity|
+      indices = entity[:indices]
+      result << text[index...indices.first]
+      result << yield(entity)
+      indices.last
+    end
+
+    result << text[last_index..-1]
+
+    Sanitize.fragment(result, Sanitize::Config::MASTODON_OUTGOING)
+  end
+
+  def format_markdown(html)
+    html = markdown_formatter.render(html)
+    html.delete("\r").delete("\n")
+  end
+
+  def markdown_formatter
+    extensions = {
+      autolink: true,
+      no_intra_emphasis: true,
+      fenced_code_blocks: true,
+      disable_indented_code_blocks: true,
+      strikethrough: true,
+      lax_spacing: true,
+      space_after_headers: true,
+      superscript: true,
+      underline: true,
+      highlight: true,
+      footnotes: false,
+    }
+
+    renderer = HTMLRenderer.new({
+      filter_html: false,
+      escape_html: false,
+      no_images: true,
+      no_styles: true,
+      safe_links_only: true,
+      hard_wrap: true,
+      link_attributes: { target: '_blank', rel: 'nofollow noopener' },
+    }) do |url|
+      link_to_url({ url: url })
+    end
+
+    Redcarpet::Markdown.new(renderer, extensions)
+  end
+end
diff --git a/app/lib/html_aware_formatter.rb b/app/lib/html_aware_formatter.rb
index 64edba09b..7a1cd0340 100644
--- a/app/lib/html_aware_formatter.rb
+++ b/app/lib/html_aware_formatter.rb
@@ -33,6 +33,10 @@ class HtmlAwareFormatter
   end
 
   def linkify
-    TextFormatter.new(text, options).to_s
+    if %w(text/markdown text/html).include?(@options[:content_type])
+      AdvancedTextFormatter.new(text, options).to_s
+    else
+      TextFormatter.new(text, options).to_s
+    end
   end
 end
diff --git a/lib/sanitize_ext/sanitize_config.rb b/lib/sanitize_ext/sanitize_config.rb
index ecaec2f84..935e1f4f6 100644
--- a/lib/sanitize_ext/sanitize_config.rb
+++ b/lib/sanitize_ext/sanitize_config.rb
@@ -55,18 +55,6 @@ class Sanitize
       end
     end
 
-    LINK_REL_TRANSFORMER = lambda do |env|
-      return unless env[:node_name] == 'a' and env[:node]['href']
-
-      node = env[:node]
-
-      rel = (node['rel'] || '').split(' ') & ['tag']
-      unless env[:config][:outgoing] && TagManager.instance.local_url?(node['href'])
-        rel += ['nofollow', 'noopener', 'noreferrer']
-      end
-      node['rel'] = rel.join(' ')
-    end
-
     UNSUPPORTED_HREF_TRANSFORMER = lambda do |env|
       return unless env[:node_name] == 'a'
 
@@ -97,6 +85,7 @@ class Sanitize
 
       add_attributes: {
         'a' => {
+          'rel' => 'nofollow noopener noreferrer',
           'target' => '_blank',
         },
       },
@@ -110,7 +99,6 @@ class Sanitize
         CLASS_WHITELIST_TRANSFORMER,
         IMG_TAG_TRANSFORMER,
         UNSUPPORTED_HREF_TRANSFORMER,
-        LINK_REL_TRANSFORMER,
       ]
     )
 
@@ -135,5 +123,48 @@ class Sanitize
         'source' => { 'src' => HTTP_PROTOCOLS }
       )
     )
+
+    LINK_REL_TRANSFORMER = lambda do |env|
+      return unless env[:node_name] == 'a' && env[:node]['href']
+
+      node = env[:node]
+
+      rel = (node['rel'] || '').split(' ') & ['tag']
+      rel += ['nofollow', 'noopener', 'noreferrer'] unless TagManager.instance.local_url?(node['href'])
+
+      if rel.empty?
+        node['rel']&.delete
+      else
+        node['rel'] = rel.join(' ')
+      end
+    end
+
+    LINK_TARGET_TRANSFORMER = lambda do |env|
+      return unless env[:node_name] == 'a' && env[:node]['href']
+
+      node = env[:node]
+      if node['target'] != '_blank' && TagManager.instance.local_url?(node['href'])
+        node['target']&.delete
+      else
+        node['target'] = '_blank'
+      end
+    end
+
+    MASTODON_OUTGOING ||= freeze_config MASTODON_STRICT.merge(
+      attributes: merge(
+        MASTODON_STRICT[:attributes],
+        'a' => %w(href rel class title target)
+      ),
+
+      add_attributes: {},
+
+      transformers: [
+        CLASS_WHITELIST_TRANSFORMER,
+        IMG_TAG_TRANSFORMER,
+        UNSUPPORTED_HREF_TRANSFORMER,
+        LINK_REL_TRANSFORMER,
+        LINK_TARGET_TRANSFORMER,
+      ]
+    )
   end
 end
diff --git a/spec/lib/advanced_text_formatter_spec.rb b/spec/lib/advanced_text_formatter_spec.rb
new file mode 100644
index 000000000..c097b86e1
--- /dev/null
+++ b/spec/lib/advanced_text_formatter_spec.rb
@@ -0,0 +1,274 @@
+require 'rails_helper'
+
+RSpec.describe AdvancedTextFormatter do
+  describe '#to_s' do
+    let(:preloaded_accounts) { nil }
+    let(:content_type) { 'text/markdown' }
+
+    subject { described_class.new(text, preloaded_accounts: preloaded_accounts, content_type: content_type).to_s }
+
+    context 'given a markdown source' do
+      let(:content_type) { 'text/markdown' }
+
+      context 'given text containing plain text' do
+        let(:text) { 'text' }
+
+        it 'paragraphizes the text' do
+          is_expected.to eq '<p>text</p>'
+        end
+      end
+
+      context 'given text containing line feeds' do
+        let(:text) { "line\nfeed" }
+
+        it 'removes line feeds' do
+          is_expected.not_to include "\n"
+        end
+      end
+
+      context 'given some inline code using backticks' do
+        let(:text) { 'test `foo` bar' }
+
+        it 'formats code using <code>' do
+          is_expected.to include 'test <code>foo</code> bar'
+        end
+      end
+
+      context 'given some quote' do
+        let(:text) { "> foo\n\nbar" }
+
+        it 'formats code using <code>' do
+          is_expected.to include '<blockquote><p>foo</p></blockquote>'
+        end
+      end
+
+      context 'given text containing linkable mentions' do
+        let(:preloaded_accounts) { [Fabricate(:account, username: 'alice')] }
+        let(:text) { '@alice' }
+
+        it 'creates a mention link' do
+          is_expected.to include '<a href="https://cb6e6126.ngrok.io/@alice" class="u-url mention">@<span>alice</span></a></span>'
+        end
+      end
+
+      context 'given text containing unlinkable mentions' do
+        let(:preloaded_accounts) { [] }
+        let(:text) { '@alice' }
+
+        it 'does not create a mention link' do
+          is_expected.to include '@alice'
+        end
+      end
+
+      context 'given a stand-alone medium URL' do
+        let(:text) { 'https://hackernoon.com/the-power-to-build-communities-a-response-to-mark-zuckerberg-3f2cac9148a4' }
+
+        it 'matches the full URL' do
+          is_expected.to include 'href="https://hackernoon.com/the-power-to-build-communities-a-response-to-mark-zuckerberg-3f2cac9148a4"'
+        end
+      end
+
+      context 'given a stand-alone google URL' do
+        let(:text) { 'http://google.com' }
+
+        it 'matches the full URL' do
+          is_expected.to include 'href="http://google.com"'
+        end
+      end
+
+      context 'given a stand-alone URL with a newer TLD' do
+        let(:text) { 'http://example.gay' }
+
+        it 'matches the full URL' do
+          is_expected.to include 'href="http://example.gay"'
+        end
+      end
+
+      context 'given a stand-alone IDN URL' do
+        let(:text) { 'https://nic.みんな/' }
+
+        it 'matches the full URL' do
+          is_expected.to include 'href="https://nic.みんな/"'
+        end
+
+        it 'has display URL' do
+          is_expected.to include '<span class="">nic.みんな/</span>'
+        end
+      end
+
+      context 'given a URL with a trailing period' do
+        let(:text) { 'http://www.mcmansionhell.com/post/156408871451/50-states-of-mcmansion-hell-scottsdale-arizona. ' }
+
+        it 'matches the full URL but not the period' do
+          is_expected.to include 'href="http://www.mcmansionhell.com/post/156408871451/50-states-of-mcmansion-hell-scottsdale-arizona"'
+        end
+      end
+
+      context 'given a URL enclosed with parentheses' do
+        let(:text) { '(http://google.com/)' }
+
+        it 'matches the full URL but not the parentheses' do
+          is_expected.to include 'href="http://google.com/"'
+        end
+      end
+
+      context 'given a URL with a trailing exclamation point' do
+        let(:text) { 'http://www.google.com!' }
+
+        it 'matches the full URL but not the exclamation point' do
+          is_expected.to include 'href="http://www.google.com"'
+        end
+      end
+
+      context 'given a URL with a trailing single quote' do
+        let(:text) { "http://www.google.com'" }
+
+        it 'matches the full URL but not the single quote' do
+          is_expected.to include 'href="http://www.google.com"'
+        end
+      end
+    end
+
+    context 'given a URL with a trailing angle bracket' do
+      let(:text) { 'http://www.google.com>' }
+
+      it 'matches the full URL but not the angle bracket' do
+        is_expected.to include 'href="http://www.google.com"'
+      end
+    end
+
+    context 'given a URL with a query string' do
+      context 'with escaped unicode character' do
+        let(:text) { 'https://www.ruby-toolbox.com/search?utf8=%E2%9C%93&q=autolink' }
+
+        it 'matches the full URL' do
+          is_expected.to include 'href="https://www.ruby-toolbox.com/search?utf8=%E2%9C%93&amp;q=autolink"'
+        end
+      end
+
+      context 'with unicode character' do
+        let(:text) { 'https://www.ruby-toolbox.com/search?utf8=✓&q=autolink' }
+
+        it 'matches the full URL' do
+          is_expected.to include 'href="https://www.ruby-toolbox.com/search?utf8=✓&amp;q=autolink"'
+        end
+      end
+
+      context 'with unicode character at the end' do
+        let(:text) { 'https://www.ruby-toolbox.com/search?utf8=✓' }
+
+        it 'matches the full URL' do
+          is_expected.to include 'href="https://www.ruby-toolbox.com/search?utf8=✓"'
+        end
+      end
+
+      context 'with escaped and not escaped unicode characters' do
+        let(:text) { 'https://www.ruby-toolbox.com/search?utf8=%E2%9C%93&utf81=✓&q=autolink' }
+
+        it 'preserves escaped unicode characters' do
+          is_expected.to include 'href="https://www.ruby-toolbox.com/search?utf8=%E2%9C%93&amp;utf81=✓&amp;q=autolink"'
+        end
+      end
+
+      context 'given a URL with parentheses in it' do
+        let(:text) { 'https://en.wikipedia.org/wiki/Diaspora_(software)' }
+
+        it 'matches the full URL' do
+          is_expected.to include 'href="https://en.wikipedia.org/wiki/Diaspora_(software)"'
+        end
+      end
+
+      context 'given a URL in quotation marks' do
+        let(:text) { '"https://example.com/"' }
+
+        it 'does not match the quotation marks' do
+          is_expected.to include 'href="https://example.com/"'
+        end
+      end
+
+      context 'given a URL in angle brackets' do
+        let(:text) { '<https://example.com/>' }
+
+        it 'does not match the angle brackets' do
+          is_expected.to include 'href="https://example.com/"'
+        end
+      end
+
+      context 'given a URL containing unsafe code (XSS attack, invisible part)' do
+        let(:text) { %q{http://example.com/blahblahblahblah/a<script>alert("Hello")</script>} }
+
+        it 'does not include the HTML in the URL' do
+          is_expected.to include '"http://example.com/blahblahblahblah/a"'
+        end
+
+        it 'does not include a script tag' do
+          is_expected.to_not include '<script>'
+        end
+      end
+
+      context 'given text containing HTML code (script tag)' do
+        let(:text) { '<script>alert("Hello")</script>' }
+
+        it 'does not include a script tag' do
+          is_expected.to_not include '<script>'
+        end
+      end
+
+      context 'given text containing HTML (XSS attack)' do
+        let(:text) { %q{<img src="javascript:alert('XSS');">} }
+
+        it 'does not include the javascript' do
+          is_expected.to_not include 'href="javascript:'
+        end
+      end
+
+      context 'given an invalid URL' do
+        let(:text) { 'http://www\.google\.com' }
+
+        it 'outputs the raw URL' do
+          is_expected.to eq '<p>http://www\.google\.com</p>'
+        end
+      end
+
+      context 'given text containing a hashtag' do
+        let(:text)  { '#hashtag' }
+
+        it 'creates a hashtag link' do
+          is_expected.to include '/tags/hashtag" class="mention hashtag" rel="tag">#<span>hashtag</span></a>'
+        end
+      end
+
+      context 'given text containing a hashtag with Unicode chars' do
+        let(:text)  { '#hashtagタグ' }
+
+        it 'creates a hashtag link' do
+          is_expected.to include '/tags/hashtag%E3%82%BF%E3%82%B0" class="mention hashtag" rel="tag">#<span>hashtagタグ</span></a>'
+        end
+      end
+
+      context 'given text with a stand-alone xmpp: URI' do
+        let(:text) { 'xmpp:user@instance.com' }
+
+        it 'matches the full URI' do
+          is_expected.to include 'href="xmpp:user@instance.com"'
+        end
+      end
+
+      context 'given text with an xmpp: URI with a query-string' do
+        let(:text) { 'please join xmpp:muc@instance.com?join right now' }
+
+        it 'matches the full URI' do
+          is_expected.to include 'href="xmpp:muc@instance.com?join"'
+        end
+      end
+
+      context 'given text containing a magnet: URI' do
+        let(:text) { 'wikipedia gives this example of a magnet uri: magnet:?xt=urn:btih:c12fe1c06bba254a9dc9f519b335aa7c1367a88a' }
+
+        it 'matches the full URI' do
+          is_expected.to include 'href="magnet:?xt=urn:btih:c12fe1c06bba254a9dc9f519b335aa7c1367a88a"'
+        end
+      end
+    end
+  end
+end
diff --git a/spec/lib/sanitize_config_spec.rb b/spec/lib/sanitize_config_spec.rb
index 8bcffb2e5..dc6418e5b 100644
--- a/spec/lib/sanitize_config_spec.rb
+++ b/spec/lib/sanitize_config_spec.rb
@@ -41,18 +41,8 @@ describe Sanitize::Config do
     end
   end
 
-  describe '::MASTODON_STRICT' do
-    subject { Sanitize::Config::MASTODON_STRICT }
-
-    it_behaves_like 'common HTML sanitization'
-
-    it 'keeps a with href and rel tag' do
-      expect(Sanitize.fragment('<a href="http://example.com" rel="tag">Test</a>', subject)).to eq '<a href="http://example.com" rel="tag nofollow noopener noreferrer" target="_blank">Test</a>'
-    end
-  end
-
-  describe '::MASTODON_STRICT with outgoing toots' do
-    subject { Sanitize::Config::MASTODON_STRICT.merge(outgoing: true) }
+  describe '::MASTODON_OUTGOING' do
+    subject { Sanitize::Config::MASTODON_OUTGOING }
 
     around do |example|
       original_web_domain = Rails.configuration.x.web_domain
@@ -62,9 +52,9 @@ describe Sanitize::Config do
 
     it_behaves_like 'common HTML sanitization'
 
-    it 'keeps a with href and rel tag, not adding to rel if url is local' do
+    it 'keeps a with href and rel tag, not adding to rel or target if url is local' do
       Rails.configuration.x.web_domain = 'domain.test'
-      expect(Sanitize.fragment('<a href="http://domain.test/tags/foo" rel="tag">Test</a>', subject)).to eq '<a href="http://domain.test/tags/foo" rel="tag" target="_blank">Test</a>'
+      expect(Sanitize.fragment('<a href="http://domain.test/tags/foo" rel="tag">Test</a>', subject)).to eq '<a href="http://domain.test/tags/foo" rel="tag">Test</a>'
     end
   end
 end