Custom emoji (#4988)

* Custom emoji

- In OStatus: `<link rel="emoji" name="coolcat" href="http://..." />`
- In ActivityPub: `{ type: "Emoji", name: ":coolcat:", href: "http://..." }`
- In REST API: Status object includes `emojis` array (`shortcode`, `url`)
- Domain blocks with reject media stop emojis
- Emoji file up to 50KB
- Web UI handles custom emojis
- Static pages render custom emojis as `<img />` tags

Side effects:

- Undo #4500 optimization, as I needed to modify it to restore
  shortcode handling in emojify()
- Formatter#plaintext should now make sure stripped out line-breaks
  and paragraphs are replaced with newlines

* Fix emoji at the start not being converted

12 files changed, 209 insertions, 28 deletions

diff --git a/app/javascript/mastodon/emoji.js b/app/javascript/mastodon/emoji.js
index a41dfdd1d..865b85b61 100644
--- a/app/javascript/mastodon/emoji.js
+++ b/app/javascript/mastodon/emoji.js
@@ -3,28 +3,48 @@ import Trie from 'substring-trie';
 
 const trie = new Trie(Object.keys(unicodeMapping));
 
-const emojify = str => {
-  let rtn = '';
-  for (;;) {
-    let match, i = 0;
-    while (i < str.length && str[i] !== '<' && !(match = trie.search(str.slice(i)))) {
-      i += str.codePointAt(i) < 65536 ? 1 : 2;
-    }
-    if (i === str.length)
-      break;
-    else if (str[i] === '<') {
-      let tagend = str.indexOf('>', i + 1) + 1;
-      if (!tagend)
-        break;
-      rtn += str.slice(0, tagend);
-      str = str.slice(tagend);
-    } else {
-      const [filename, shortCode] = unicodeMapping[match];
-      rtn += str.slice(0, i) + `<img draggable="false" class="emojione" alt="${match}" title=":${shortCode}:" src="/emoji/${filename}.svg" />`;
-      str = str.slice(i + match.length);
+const emojify = (str, customEmojis = {}) => {
+  // This walks through the string from start to end, ignoring any tags (<p>, <br>, etc.)
+  // and replacing valid unicode strings
+  // that _aren't_ within tags with an <img> version.
+  // The goal is to be the same as an emojione.regUnicode replacement, but faster.
+  let i = -1;
+  let insideTag = false;
+  let insideShortname = false;
+  let shortnameStartIndex = -1;
+  let match;
+  while (++i < str.length) {
+    const char = str.charAt(i);
+    if (insideShortname && char === ':') {
+      const shortname = str.substring(shortnameStartIndex, i + 1);
+      if (shortname in customEmojis) {
+        const replacement = `<img draggable="false" class="emojione" alt="${shortname}" title="${shortname}" src="${customEmojis[shortname]}" />`;
+        str = str.substring(0, shortnameStartIndex) + replacement + str.substring(i + 1);
+        i += (replacement.length - shortname.length - 1); // jump ahead the length we've added to the string
+      } else {
+        i--;
+      }
+      insideShortname = false;
+    } else if (insideTag && char === '>') {
+      insideTag = false;
+    } else if (char === '<') {
+      insideTag = true;
+      insideShortname = false;
+    } else if (!insideTag && char === ':') {
+      insideShortname = true;
+      shortnameStartIndex = i;
+    } else if (!insideTag && (match = trie.search(str.substring(i)))) {
+      const unicodeStr = match;
+      if (unicodeStr in unicodeMapping) {
+        const [filename, shortCode] = unicodeMapping[unicodeStr];
+        const alt      = unicodeStr;
+        const replacement =  `<img draggable="false" class="emojione" alt="${alt}" title=":${shortCode}:" src="/emoji/${filename}.svg" />`;
+        str = str.substring(0, i) + replacement + str.substring(i + unicodeStr.length);
+        i += (replacement.length - unicodeStr.length); // jump ahead the length we've added to the string
+      }
     }
   }
-  return rtn + str;
+  return str;
 };
 
 export default emojify;
diff --git a/app/javascript/mastodon/reducers/statuses.js b/app/javascript/mastodon/reducers/statuses.js
index 7f906bef6..38b23504e 100644
--- a/app/javascript/mastodon/reducers/statuses.js
+++ b/app/javascript/mastodon/reducers/statuses.js
@@ -58,9 +58,14 @@ const normalizeStatus = (state, status) => {
   }
 
   const searchContent = [status.spoiler_text, status.content].join(' ').replace(/<br \/>/g, '\n').replace(/<\/p><p>/g, '\n\n');
+  const emojiMap = normalStatus.emojis.reduce((obj, emoji) => {
+    obj[`:${emoji.shortcode}:`] = emoji.url;
+    return obj;
+  }, {});
+
   normalStatus.search_index = domParser.parseFromString(searchContent, 'text/html').documentElement.textContent;
-  normalStatus.contentHtml = emojify(normalStatus.content);
-  normalStatus.spoilerHtml = emojify(escapeTextContentForBrowser(normalStatus.spoiler_text || ''));
+  normalStatus.contentHtml = emojify(normalStatus.content, emojiMap);
+  normalStatus.spoilerHtml = emojify(escapeTextContentForBrowser(normalStatus.spoiler_text || ''), emojiMap);
 
   return state.update(status.id, ImmutableMap(), map => map.mergeDeep(fromJS(normalStatus)));
 };
diff --git a/app/lib/activitypub/activity/create.rb b/app/lib/activitypub/activity/create.rb
index 894759d9a..41f2b0bad 100644
--- a/app/lib/activitypub/activity/create.rb
+++ b/app/lib/activitypub/activity/create.rb
@@ -61,6 +61,8 @@ class ActivityPub::Activity::Create < ActivityPub::Activity
         process_hashtag tag, status
       when 'Mention'
         process_mention tag, status
+      when 'Emoji'
+        process_emoji tag, status
       end
     end
   end
@@ -79,6 +81,17 @@ class ActivityPub::Activity::Create < ActivityPub::Activity
     account.mentions.create(status: status)
   end
 
+  def process_emoji(tag, _status)
+    shortcode = tag['name'].delete(':')
+    emoji     = CustomEmoji.find_by(shortcode: shortcode, domain: @account.domain)
+
+    return if !emoji.nil? || skip_download?
+
+    emoji = CustomEmoji.new(domain: @account.domain, shortcode: shortcode)
+    emoji.image_remote_url = tag['href']
+    emoji.save
+  end
+
   def process_attachments(status)
     return unless @object['attachment'].is_a?(Array)
 
diff --git a/app/lib/formatter.rb b/app/lib/formatter.rb
index 575830190..29fea27de 100644
--- a/app/lib/formatter.rb
+++ b/app/lib/formatter.rb
@@ -9,7 +9,7 @@ class Formatter
 
   include ActionView::Helpers::TextHelper
 
-  def format(status)
+  def format(status, options = {})
     if status.reblog?
       prepend_reblog = status.reblog.account.acct
       status         = status.proper
@@ -19,7 +19,11 @@ class Formatter
 
     raw_content = status.text
 
-    return reformat(raw_content) unless status.local?
+    unless status.local?
+      html = reformat(raw_content)
+      html = encode_custom_emojis(html, status.emojis) if options[:custom_emojify]
+      return html
+    end
 
     linkable_accounts = status.mentions.map(&:account)
     linkable_accounts << status.account
@@ -27,6 +31,7 @@ class Formatter
     html = raw_content
     html = "RT @#{prepend_reblog} #{html}" if prepend_reblog
     html = encode_and_link_urls(html, linkable_accounts)
+    html = encode_custom_emojis(html, status.emojis) if options[:custom_emojify]
     html = simple_format(html, {}, sanitize: false)
     html = html.delete("\n")
 
@@ -39,7 +44,9 @@ class Formatter
 
   def plaintext(status)
     return status.text if status.local?
-    strip_tags(status.text)
+
+    text = status.text.gsub(/(<br \/>|<br>|<\/p>)+/) { |match| "#{match}\n" }
+    strip_tags(text)
   end
 
   def simplified_format(account)
@@ -76,6 +83,47 @@ class Formatter
     end
   end
 
+  def encode_custom_emojis(html, emojis)
+    return html if emojis.empty?
+
+    emoji_map = emojis.map { |e| [e.shortcode, full_asset_url(e.image.url)] }.to_h
+
+    i                     = -1
+    inside_tag            = false
+    inside_shortname      = false
+    shortname_start_index = -1
+
+    while i + 1 < html.size
+      i += 1
+
+      if inside_shortname && html[i] == ':'
+        shortcode = html[shortname_start_index + 1..i - 1]
+        emoji     = emoji_map[shortcode]
+
+        if emoji
+          replacement = "<img draggable=\"false\" class=\"emojione\" alt=\":#{shortcode}:\" title=\":#{shortcode}:\" src=\"#{emoji}\" />"
+          before_html = shortname_start_index.positive? ? html[0..shortname_start_index - 1] : ''
+          html        = before_html + replacement + html[i + 1..-1]
+          i          += replacement.size - (shortcode.size + 2) - 1
+        else
+          i -= 1
+        end
+
+        inside_shortname = false
+      elsif inside_tag && html[i] == '>'
+        inside_tag = false
+      elsif html[i] == '<'
+        inside_tag       = true
+        inside_shortname = false
+      elsif !inside_tag && html[i] == ':'
+        inside_shortname      = true
+        shortname_start_index = i
+      end
+    end
+
+    html
+  end
+
   def rewrite(text, entities)
     chars = text.to_s.to_char_a
 
diff --git a/app/lib/ostatus/activity/creation.rb b/app/lib/ostatus/activity/creation.rb
index 1a23c9efa..d3f1629c4 100644
--- a/app/lib/ostatus/activity/creation.rb
+++ b/app/lib/ostatus/activity/creation.rb
@@ -42,6 +42,7 @@ class OStatus::Activity::Creation < OStatus::Activity::Base
       save_mentions(status)
       save_hashtags(status)
       save_media(status)
+      save_emojis(status)
     end
 
     if thread? && status.thread.nil?
@@ -150,6 +151,25 @@ class OStatus::Activity::Creation < OStatus::Activity::Base
     end
   end
 
+  def save_emojis(parent)
+    do_not_download = DomainBlock.find_by(domain: parent.account.domain)&.reject_media?
+
+    return if do_not_download
+
+    @xml.xpath('./xmlns:link[@rel="emoji"]', xmlns: TagManager::XMLNS).each do |link|
+      next unless link['href'] && link['name']
+
+      shortcode = link['name'].delete(':')
+      emoji     = CustomEmoji.find_by(shortcode: shortcode, domain: parent.account.domain)
+
+      next unless emoji.nil?
+
+      emoji = CustomEmoji.new(shortcode: shortcode, domain: parent.account.domain)
+      emoji.image_remote_url = link['href']
+      emoji.save
+    end
+  end
+
   def account_from_href(href)
     url = Addressable::URI.parse(href).normalize
 
diff --git a/app/lib/ostatus/atom_serializer.rb b/app/lib/ostatus/atom_serializer.rb
index b8e22a381..a6a5cb0c4 100644
--- a/app/lib/ostatus/atom_serializer.rb
+++ b/app/lib/ostatus/atom_serializer.rb
@@ -368,5 +368,9 @@ class OStatus::AtomSerializer
     end
 
     append_element(entry, 'mastodon:scope', status.visibility)
+
+    status.emojis.each do |emoji|
+      append_element(entry, 'link', nil, rel: :emoji, href: full_asset_url(emoji.image.url), name: emoji.shortcode)
+    end
   end
 end
diff --git a/app/models/custom_emoji.rb b/app/models/custom_emoji.rb
new file mode 100644
index 000000000..f4d3b16a0
--- /dev/null
+++ b/app/models/custom_emoji.rb
@@ -0,0 +1,38 @@
+# frozen_string_literal: true
+# == Schema Information
+#
+# Table name: custom_emojis
+#
+#  id                 :integer          not null, primary key
+#  shortcode          :string           default(""), not null
+#  domain             :string
+#  image_file_name    :string
+#  image_content_type :string
+#  image_file_size    :integer
+#  image_updated_at   :datetime
+#  created_at         :datetime         not null
+#  updated_at         :datetime         not null
+#
+
+class CustomEmoji < ApplicationRecord
+  SHORTCODE_RE_FRAGMENT = '[a-zA-Z0-9_]{2,}'
+
+  SCAN_RE = /(?<=[^[:alnum:]:]|\n|^)
+    :(#{SHORTCODE_RE_FRAGMENT}):
+    (?=[^[:alnum:]:]|$)/x
+
+  has_attached_file :image
+
+  validates_attachment :image, content_type: { content_type: 'image/png' }, presence: true, size: { in: 0..50.kilobytes }
+  validates :shortcode, uniqueness: { scope: :domain }, format: { with: /\A#{SHORTCODE_RE_FRAGMENT}\z/ }, length: { minimum: 2 }
+
+  include Remotable
+
+  class << self
+    def from_text(text, domain)
+      return [] if text.blank?
+      shortcodes = text.scan(SCAN_RE).map(&:first)
+      where(shortcode: shortcodes, domain: domain)
+    end
+  end
+end
diff --git a/app/models/status.rb b/app/models/status.rb
index 2a2cdcf6e..326d128d6 100644
--- a/app/models/status.rb
+++ b/app/models/status.rb
@@ -131,6 +131,10 @@ class Status < ApplicationRecord
     !sensitive? && media_attachments.any?
   end
 
+  def emojis
+    CustomEmoji.from_text(text, account.domain)
+  end
+
   after_create :store_uri, if: :local?
 
   before_validation :prepare_contents, if: :local?
diff --git a/app/serializers/activitypub/note_serializer.rb b/app/serializers/activitypub/note_serializer.rb
index 166214eee..e5d8e3f03 100644
--- a/app/serializers/activitypub/note_serializer.rb
+++ b/app/serializers/activitypub/note_serializer.rb
@@ -57,7 +57,7 @@ class ActivityPub::NoteSerializer < ActiveModel::Serializer
   end
 
   def virtual_tags
-    object.mentions + object.tags
+    object.mentions + object.tags + object.emojis
   end
 
   def atom_uri
@@ -137,4 +137,22 @@ class ActivityPub::NoteSerializer < ActiveModel::Serializer
       "##{object.name}"
     end
   end
+
+  class CustomEmojiSerializer < ActiveModel::Serializer
+    include RoutingHelper
+
+    attributes :type, :href, :name
+
+    def type
+      'Emoji'
+    end
+
+    def href
+      full_asset_url(object.image.url)
+    end
+
+    def name
+      ":#{object.shortcode}:"
+    end
+  end
 end
diff --git a/app/serializers/rest/status_serializer.rb b/app/serializers/rest/status_serializer.rb
index 298a3bb40..d8efa8e60 100644
--- a/app/serializers/rest/status_serializer.rb
+++ b/app/serializers/rest/status_serializer.rb
@@ -17,6 +17,7 @@ class REST::StatusSerializer < ActiveModel::Serializer
   has_many :media_attachments, serializer: REST::MediaAttachmentSerializer
   has_many :mentions
   has_many :tags
+  has_many :emojis
 
   def current_user?
     !current_user.nil?
@@ -106,4 +107,14 @@ class REST::StatusSerializer < ActiveModel::Serializer
       tag_url(object)
     end
   end
+
+  class CustomEmojiSerializer < ActiveModel::Serializer
+    include RoutingHelper
+
+    attributes :shortcode, :url
+
+    def url
+      full_asset_url(object.image.url)
+    end
+  end
 end
diff --git a/app/views/stream_entries/_detailed_status.html.haml b/app/views/stream_entries/_detailed_status.html.haml
index dd9456260..692d5a6d5 100644
--- a/app/views/stream_entries/_detailed_status.html.haml
+++ b/app/views/stream_entries/_detailed_status.html.haml
@@ -17,7 +17,7 @@
       %p{ style: 'margin-bottom: 0' }<
         %span.p-summary> #{status.spoiler_text}&nbsp;
         %a.status__content__spoiler-link{ href: '#' }= t('statuses.show_more')
-    .e-content{ lang: status.language, style: "display: #{status.spoiler_text? ? 'none' : 'block'}; direction: #{rtl_status?(status) ? 'rtl' : 'ltr'}" }= Formatter.instance.format(status)
+    .e-content{ lang: status.language, style: "display: #{status.spoiler_text? ? 'none' : 'block'}; direction: #{rtl_status?(status) ? 'rtl' : 'ltr'}" }= Formatter.instance.format(status, custom_emojify: true)
 
   - if !status.media_attachments.empty?
     - if status.media_attachments.first.video?
diff --git a/app/views/stream_entries/_simple_status.html.haml b/app/views/stream_entries/_simple_status.html.haml
index 55aa97f32..f9a530d38 100644
--- a/app/views/stream_entries/_simple_status.html.haml
+++ b/app/views/stream_entries/_simple_status.html.haml
@@ -18,7 +18,7 @@
       %p{ style: 'margin-bottom: 0' }<
         %span.p-summary> #{status.spoiler_text}&nbsp;
         %a.status__content__spoiler-link{ href: '#' }= t('statuses.show_more')
-    .e-content{ lang: status.language, style: "display: #{status.spoiler_text? ? 'none' : 'block'}; direction: #{rtl_status?(status) ? 'rtl' : 'ltr'}" }= Formatter.instance.format(status)
+    .e-content{ lang: status.language, style: "display: #{status.spoiler_text? ? 'none' : 'block'}; direction: #{rtl_status?(status) ? 'rtl' : 'ltr'}" }= Formatter.instance.format(status, custom_emojify: true)
 
   - unless status.media_attachments.empty?
     - if status.media_attachments.first.video?