# frozen_string_literal: true require 'singleton' require_relative './sanitize_config' class HTMLRenderer < Redcarpet::Render::HTML def block_code(code, language) "
#{encode(code).gsub("\n", "
")}
"
end
def autolink(link, link_type)
return link if link_type == :email
Formatter.instance.link_url(link)
end
private
def html_entities
@html_entities ||= HTMLEntities.new
end
def encode(html)
html_entities.encode(html)
end
end
# rubocop:disable Metrics/ClassLength
class Formatter
include Singleton
include RoutingHelper
include ActionView::Helpers::TextHelper
def format(status, **options)
if status.reblog?
prepend_reblog = status.reblog.account.acct
status = status.proper
else
prepend_reblog = false
end
summary = nil
raw_content = status.text
summary_mode = false
if status.title.present?
summary = status.spoiler_text.presence || status.text
summary_mode = !options[:article_content]
raw_content = summary_mode ? summary : status.text
end
if options[:inline_poll_options] && status.preloadable_poll
raw_content = raw_content + "\n\n" + status.preloadable_poll.options.map { |title| "[ ] #{title}" }.join("\n")
end
return '' if raw_content.blank?
return format_remote_content(raw_content, status.emojis, summary: summary, **options) unless status.local?
html = raw_content
html = "RT @#{prepend_reblog} #{html}" if prepend_reblog
html = "📄 #{html}" if summary_mode
return html if options[:plaintext]
linkable_accounts = status.mentions.map(&:account)
linkable_accounts << status.account
keep_html = !summary_mode && %w(text/markdown text/html).include?(status.content_type)
html = format_markdown(html) if !summary_mode && status.content_type == 'text/markdown'
html = encode_and_link_urls(html, linkable_accounts, keep_html: keep_html)
html = reformat(html, true) if keep_html
html = encode_custom_emojis(html, status.emojis, options[:autoplay]) if options[:custom_emojify]
unless keep_html
html = simple_format(html, {}, sanitize: false)
html.delete!("\n")
end
html = summary_mode ? format_article_summary(html, status) : format_article_content(summary, html) if summary.present?
html = format_footer(html, status.footer, linkable_accounts, status.emojis, **options) if status.footer.present?
html.html_safe # rubocop:disable Rails/OutputSafety
end
def format_remote_content(html, emojis, **options)
html = reformat(html, options[:outgoing])
html = encode_custom_emojis(html, emojis, options[:autoplay]) if options[:custom_emojify]
html = format_article_content(options[:summary], html) if options[:article_content] && options[:summary].present?
html.html_safe # rubocop:disable Rails/OutputSafety
end
def format_footer(html, footer, linkable_accounts, emojis, **options)
footer = encode_and_link_urls(footer, linkable_accounts)
footer = encode_custom_emojis(footer, emojis, options[:autoplay]) if options[:custom_emojify]
footer = "– #{footer}"
footer = simple_format(footer, { 'data-name': 'footer' }, sanitize: false)
footer.delete!("\n")
"#{html}#{footer}"
end
def format_markdown(html)
html = markdown_formatter.render(html)
html.delete("\r").delete("\n")
end
def format_article(text)
text = text.gsub(/>[\r\n]+, '><')
text.html_safe # rubocop:disable Rails/OutputSafety
end
def format_article_summary(html, status)
status_url = ActivityPub::TagManager.instance.url_for(status)
"#{html}\n#{link_url(status_url)}
" end def format_article_content(summary, html) "#{format_summary(summary, html)}#{html}" end def reformat(html, outgoing = false) sanitize(html, Sanitize::Config::MASTODON_STRICT.merge(outgoing: outgoing)) rescue ArgumentError '' end def plaintext(status) return status.text if status.local? text = status.text.gsub(/(
.*?<\/p>)/im].presence || '🗎❓' end def format_spoiler(status, **options) html = encode(status.title.presence || status.spoiler_text) html = encode_custom_emojis(html, status.emojis, options[:autoplay]) html.html_safe # rubocop:disable Rails/OutputSafety end def format_poll_option(status, option, **options) html = encode(option.title) html = encode_custom_emojis(html, status.emojis, options[:autoplay]) html.html_safe # rubocop:disable Rails/OutputSafety end def format_display_name(account, **options) html = encode(account.display_name.presence || account.username) html = encode_custom_emojis(html, account.emojis, options[:autoplay]) if options[:custom_emojify] html.html_safe # rubocop:disable Rails/OutputSafety end def format_field(account, str, **options) html = account.local? ? encode_and_link_urls(str, me: true) : reformat(str) html = encode_custom_emojis(html, account.emojis, options[:autoplay]) if options[:custom_emojify] html.html_safe # rubocop:disable Rails/OutputSafety end def linkify(text, accounts = nil, options = {}) html = encode_and_link_urls(text, accounts, options) html = simple_format(html, {}, sanitize: false) html = html.delete("\n") html.html_safe # rubocop:disable Rails/OutputSafety end def link_url(url) "#{link_html(url)}" end private 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: false, no_styles: true, safe_links_only: true, hard_wrap: true, link_attributes: { target: '_blank', rel: 'nofollow noopener' }, }) Redcarpet::Markdown.new(renderer, extensions) end def html_entities @html_entities ||= HTMLEntities.new end def encode(html) html_entities.encode(html) end def encode_and_link_urls(html, accounts = nil, options = {}) if accounts.is_a?(Hash) options = accounts accounts = nil end entities = options[:keep_html] ? html_friendly_extractor(html) : utf8_friendly_extractor(html, extract_url_without_protocol: false) rewrite(html.dup, entities, options[:keep_html]) do |entity| if entity[:url] link_to_url(entity, options) elsif entity[:hashtag] link_to_hashtag(entity) elsif entity[:screen_name] link_to_mention(entity, accounts) end end end def count_tag_nesting(tag) if tag[1] == '/' then -1 elsif tag[-2] == '/' then 0 else 1 end end def encode_custom_emojis(html, emojis, animate = false) return html if emojis.empty? emoji_map = emojis.each_with_object({}) { |e, h| h[e.shortcode] = [full_asset_url(e.image.url), full_asset_url(e.image.url(:static))] } i = -1 tag_open_index = nil inside_shortname = false shortname_start_index = -1 invisible_depth = 0 while i + 1 < html.size i += 1 if invisible_depth.zero? && inside_shortname && html[i] == ':' shortcode = html[shortname_start_index + 1..i - 1] emoji = emoji_map[shortcode] if emoji original_url, static_url = emoji replacement = begin if animate "" else "" end end 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 tag_open_index && html[i] == '>' tag = html[tag_open_index..i] tag_open_index = nil if invisible_depth.positive? invisible_depth += count_tag_nesting(tag) elsif tag == '' invisible_depth = 1 end elsif html[i] == '<' tag_open_index = i inside_shortname = false elsif !tag_open_index && html[i] == ':' inside_shortname = true shortname_start_index = i end end html end def rewrite(text, entities, keep_html = false) text = text.to_s # Sort by start index entities = entities.sort_by do |entity| indices = entity.respond_to?(:indices) ? entity.indices : entity[:indices] indices.first end result = [] last_index = entities.reduce(0) do |index, entity| indices = entity.respond_to?(:indices) ? entity.indices : entity[:indices] result << (keep_html ? text[index...indices.first] : encode(text[index...indices.first])) result << yield(entity) indices.last end result << (keep_html ? text[last_index..-1] : encode(text[last_index..-1])) result.flatten.join end UNICODE_ESCAPE_BLACKLIST_RE = /\p{Z}|\p{P}/ def utf8_friendly_extractor(text, options = {}) old_to_new_index = [0] escaped = text.chars.map do |c| output = begin if c.ord.to_s(16).length > 2 && UNICODE_ESCAPE_BLACKLIST_RE.match(c).nil? CGI.escape(c) else c end end old_to_new_index << old_to_new_index.last + output.length output end.join # Note: I couldn't obtain list_slug with @user/list-name format # for mention so this requires additional check special = Extractor.extract_urls_with_indices(escaped, options).map do |extract| new_indices = [ old_to_new_index.find_index(extract[:indices].first), old_to_new_index.find_index(extract[:indices].last), ] next extract.merge( indices: new_indices, url: text[new_indices.first..new_indices.last - 1] ) end standard = Extractor.extract_entities_with_indices(text, options) extra = Extractor.extract_extra_uris_with_indices(text, options) Extractor.remove_overlapping_entities(special + standard + extra) end def html_friendly_extractor(html, options = {}) gaps = [] total_offset = 0 escaped = html.gsub(/<[^>]*>|[0-9]+;/) do |match| total_offset += match.length - 1 end_offset = Regexp.last_match.end(0) gaps << [end_offset - total_offset, total_offset] "\u200b" end entities = Extractor.extract_hashtags_with_indices(escaped, :check_url_overlap => false) + Extractor.extract_mentions_or_lists_with_indices(escaped) Extractor.remove_overlapping_entities(entities).map do |extract| pos = extract[:indices].first offset_idx = gaps.rindex { |gap| gap.first <= pos } offset = offset_idx.nil? ? 0 : gaps[offset_idx].last next extract.merge( :indices => [extract[:indices].first + offset, extract[:indices].last + offset] ) end end def link_to_url(entity, options = {}) url = Addressable::URI.parse(entity[:url]) html_attrs = { target: '_blank', rel: 'nofollow noopener noreferrer' } html_attrs[:rel] = "me #{html_attrs[:rel]}" if options[:me] Twitter::Autolink.send(:link_to_text, entity, link_html(entity[:url]), url, html_attrs) rescue Addressable::URI::InvalidURIError, IDN::Idna::IdnaError encode(entity[:url]) end def link_to_mention(entity, linkable_accounts) acct = entity[:screen_name] return link_to_account(acct) unless linkable_accounts account = linkable_accounts.find { |item| TagManager.instance.same_acct?(item.acct, acct) } account ? mention_html(account) : "@#{encode(acct)}" end def link_to_account(acct) username, domain = acct.split('@') domain = nil if TagManager.instance.local_domain?(domain) account = EntityCache.instance.mention(username, domain) account ? mention_html(account) : "@#{encode(acct)}" end def link_to_hashtag(entity) hashtag_html(entity[:hashtag]) end def link_html(url) url = Addressable::URI.parse(url).to_s prefix = url.match(/\A(https?:\/\/(www\.)?|xmpp:)/).to_s text = url[prefix.length, 30] suffix = url[prefix.length + 30..-1] cutoff = url[prefix.length..-1].length > 30 "#{encode(prefix)}#{encode(text)}#{encode(suffix)}" end def hashtag_html(tag) "##{encode(tag)}" end def mention_html(account) "@#{encode(account.username)}" end end # rubocop:enable Metrics/ClassLength