Chris@909: require 'set' Chris@909: Chris@909: module CodeRay Chris@909: module Encoders Chris@909: Chris@909: # = HTML Encoder Chris@909: # Chris@909: # This is CodeRay's most important highlighter: Chris@909: # It provides save, fast XHTML generation and CSS support. Chris@909: # Chris@909: # == Usage Chris@909: # Chris@909: # require 'coderay' Chris@909: # puts CodeRay.scan('Some /code/', :ruby).html #-> a HTML page Chris@909: # puts CodeRay.scan('Some /code/', :ruby).html(:wrap => :span) Chris@909: # #-> Some /code/ Chris@909: # puts CodeRay.scan('Some /code/', :ruby).span #-> the same Chris@909: # Chris@909: # puts CodeRay.scan('Some code', :ruby).html( Chris@909: # :wrap => nil, Chris@909: # :line_numbers => :inline, Chris@909: # :css => :style Chris@909: # ) Chris@909: # Chris@909: # == Options Chris@909: # Chris@909: # === :tab_width Chris@909: # Convert \t characters to +n+ spaces (a number.) Chris@909: # Chris@909: # Default: 8 Chris@909: # Chris@909: # === :css Chris@909: # How to include the styles; can be :class or :style. Chris@909: # Chris@909: # Default: :class Chris@909: # Chris@909: # === :wrap Chris@909: # Wrap in :page, :div, :span or nil. Chris@909: # Chris@909: # You can also use Encoders::Div and Encoders::Span. Chris@909: # Chris@909: # Default: nil Chris@909: # Chris@909: # === :title Chris@909: # Chris@909: # The title of the HTML page (works only when :wrap is set to :page.) Chris@909: # Chris@909: # Default: 'CodeRay output' Chris@909: # Chris@909: # === :line_numbers Chris@909: # Include line numbers in :table, :inline, or nil (no line numbers) Chris@909: # Chris@909: # Default: nil Chris@909: # Chris@909: # === :line_number_anchors Chris@909: # Adds anchors and links to the line numbers. Can be false (off), true (on), Chris@909: # or a prefix string that will be prepended to the anchor name. Chris@909: # Chris@909: # The prefix must consist only of letters, digits, and underscores. Chris@909: # Chris@909: # Default: true, default prefix name: "line" Chris@909: # Chris@909: # === :line_number_start Chris@909: # Where to start with line number counting. Chris@909: # Chris@909: # Default: 1 Chris@909: # Chris@909: # === :bold_every Chris@909: # Make every +n+-th number appear bold. Chris@909: # Chris@909: # Default: 10 Chris@909: # Chris@909: # === :highlight_lines Chris@909: # Chris@909: # Highlights certain line numbers. Chris@909: # Can be any Enumerable, typically just an Array or Range, of numbers. Chris@909: # Chris@909: # Bolding is deactivated when :highlight_lines is set. It only makes sense Chris@909: # in combination with :line_numbers. Chris@909: # Chris@909: # Default: nil Chris@909: # Chris@909: # === :hint Chris@909: # Include some information into the output using the title attribute. Chris@909: # Can be :info (show token kind on mouse-over), :info_long (with full path) Chris@909: # or :debug (via inspect). Chris@909: # Chris@909: # Default: false Chris@909: class HTML < Encoder Chris@909: Chris@909: register_for :html Chris@909: Chris@909: FILE_EXTENSION = 'snippet.html' Chris@909: Chris@909: DEFAULT_OPTIONS = { Chris@909: :tab_width => 8, Chris@909: Chris@909: :css => :class, Chris@909: :style => :alpha, Chris@909: :wrap => nil, Chris@909: :title => 'CodeRay output', Chris@909: Chris@909: :line_numbers => nil, Chris@909: :line_number_anchors => 'n', Chris@909: :line_number_start => 1, Chris@909: :bold_every => 10, Chris@909: :highlight_lines => nil, Chris@909: Chris@909: :hint => false, Chris@909: } Chris@909: Chris@909: autoload :Output, 'coderay/encoders/html/output' Chris@909: autoload :CSS, 'coderay/encoders/html/css' Chris@909: autoload :Numbering, 'coderay/encoders/html/numbering' Chris@909: Chris@909: attr_reader :css Chris@909: Chris@909: protected Chris@909: Chris@909: HTML_ESCAPE = { #:nodoc: Chris@909: '&' => '&', Chris@909: '"' => '"', Chris@909: '>' => '>', Chris@909: '<' => '<', Chris@909: } Chris@909: Chris@909: # This was to prevent illegal HTML. Chris@909: # Strange chars should still be avoided in codes. Chris@909: evil_chars = Array(0x00...0x20) - [?\n, ?\t, ?\s] Chris@909: evil_chars.each { |i| HTML_ESCAPE[i.chr] = ' ' } Chris@909: #ansi_chars = Array(0x7f..0xff) Chris@909: #ansi_chars.each { |i| HTML_ESCAPE[i.chr] = '&#%d;' % i } Chris@909: # \x9 (\t) and \xA (\n) not included Chris@909: #HTML_ESCAPE_PATTERN = /[\t&"><\0-\x8\xB-\x1f\x7f-\xff]/ Chris@909: HTML_ESCAPE_PATTERN = /[\t"&><\0-\x8\xB-\x1f]/ Chris@909: Chris@909: TOKEN_KIND_TO_INFO = Hash.new do |h, kind| Chris@909: h[kind] = kind.to_s.gsub(/_/, ' ').gsub(/\b\w/) { $&.capitalize } Chris@909: end Chris@909: Chris@909: TRANSPARENT_TOKEN_KINDS = Set[ Chris@909: :delimiter, :modifier, :content, :escape, :inline_delimiter, Chris@909: ] Chris@909: Chris@909: # Generate a hint about the given +kinds+ in a +hint+ style. Chris@909: # Chris@909: # +hint+ may be :info, :info_long or :debug. Chris@909: def self.token_path_to_hint hint, kinds Chris@909: kinds = Array kinds Chris@909: title = Chris@909: case hint Chris@909: when :info Chris@909: kinds = kinds[1..-1] if TRANSPARENT_TOKEN_KINDS.include? kinds.first Chris@909: TOKEN_KIND_TO_INFO[kinds.first] Chris@909: when :info_long Chris@909: kinds.reverse.map { |kind| TOKEN_KIND_TO_INFO[kind] }.join('/') Chris@909: when :debug Chris@909: kinds.inspect Chris@909: end Chris@909: title ? " title=\"#{title}\"" : '' Chris@909: end Chris@909: Chris@909: def setup options Chris@909: super Chris@909: Chris@909: if options[:wrap] || options[:line_numbers] Chris@909: @real_out = @out Chris@909: @out = '' Chris@909: end Chris@909: Chris@909: @HTML_ESCAPE = HTML_ESCAPE.dup Chris@909: @HTML_ESCAPE["\t"] = ' ' * options[:tab_width] Chris@909: Chris@909: @opened = [] Chris@909: @last_opened = nil Chris@909: @css = CSS.new options[:style] Chris@909: Chris@909: hint = options[:hint] Chris@909: if hint && ![:debug, :info, :info_long].include?(hint) Chris@909: raise ArgumentError, "Unknown value %p for :hint; \ Chris@909: expected :info, :info_long, :debug, false, or nil." % hint Chris@909: end Chris@909: Chris@909: css_classes = TokenKinds Chris@909: case options[:css] Chris@909: when :class Chris@909: @span_for_kind = Hash.new do |h, k| Chris@909: if k.is_a? ::Symbol Chris@909: kind = k_dup = k Chris@909: else Chris@909: kind = k.first Chris@909: k_dup = k.dup Chris@909: end Chris@909: if kind != :space && (hint || css_class = css_classes[kind]) Chris@909: title = HTML.token_path_to_hint hint, k if hint Chris@909: css_class ||= css_classes[kind] Chris@909: h[k_dup] = "" Chris@909: else Chris@909: h[k_dup] = nil Chris@909: end Chris@909: end Chris@909: when :style Chris@909: @span_for_kind = Hash.new do |h, k| Chris@909: kind = k.is_a?(Symbol) ? k : k.first Chris@909: h[k.is_a?(Symbol) ? k : k.dup] = Chris@909: if kind != :space && (hint || css_classes[kind]) Chris@909: title = HTML.token_path_to_hint hint, k if hint Chris@909: style = @css.get_style Array(k).map { |c| css_classes[c] } Chris@909: "" Chris@909: end Chris@909: end Chris@909: else Chris@909: raise ArgumentError, "Unknown value %p for :css." % options[:css] Chris@909: end Chris@909: Chris@909: @set_last_opened = options[:hint] || options[:css] == :style Chris@909: end Chris@909: Chris@909: def finish options Chris@909: unless @opened.empty? Chris@909: warn '%d tokens still open: %p' % [@opened.size, @opened] if $CODERAY_DEBUG Chris@909: @out << '' while @opened.pop Chris@909: @last_opened = nil Chris@909: end Chris@909: Chris@909: @out.extend Output Chris@909: @out.css = @css Chris@909: if options[:line_numbers] Chris@909: Numbering.number! @out, options[:line_numbers], options Chris@909: end Chris@909: @out.wrap! options[:wrap] Chris@909: @out.apply_title! options[:title] Chris@909: Chris@909: if defined?(@real_out) && @real_out Chris@909: @real_out << @out Chris@909: @out = @real_out Chris@909: end Chris@909: Chris@909: super Chris@909: end Chris@909: Chris@909: public Chris@909: Chris@909: def text_token text, kind Chris@909: if text =~ /#{HTML_ESCAPE_PATTERN}/o Chris@909: text = text.gsub(/#{HTML_ESCAPE_PATTERN}/o) { |m| @HTML_ESCAPE[m] } Chris@909: end Chris@909: if style = @span_for_kind[@last_opened ? [kind, *@opened] : kind] Chris@909: @out << style << text << '' Chris@909: else Chris@909: @out << text Chris@909: end Chris@909: end Chris@909: Chris@909: # token groups, eg. strings Chris@909: def begin_group kind Chris@909: @out << (@span_for_kind[@last_opened ? [kind, *@opened] : kind] || '') Chris@909: @opened << kind Chris@909: @last_opened = kind if @set_last_opened Chris@909: end Chris@909: Chris@909: def end_group kind Chris@909: if $CODERAY_DEBUG && (@opened.empty? || @opened.last != kind) Chris@909: warn 'Malformed token stream: Trying to close a token (%p) ' \ Chris@909: 'that is not open. Open are: %p.' % [kind, @opened[1..-1]] Chris@909: end Chris@909: if @opened.pop Chris@909: @out << '' Chris@909: @last_opened = @opened.last if @last_opened Chris@909: end Chris@909: end Chris@909: Chris@909: # whole lines to be highlighted, eg. a deleted line in a diff Chris@909: def begin_line kind Chris@909: if style = @span_for_kind[@last_opened ? [kind, *@opened] : kind] Chris@909: if style['class="'] Chris@909: @out << style.sub('class="', 'class="line ') Chris@909: else Chris@909: @out << style.sub('>', ' class="line">') Chris@909: end Chris@909: else Chris@909: @out << '' Chris@909: end Chris@909: @opened << kind Chris@909: @last_opened = kind if @options[:css] == :style Chris@909: end Chris@909: Chris@909: def end_line kind Chris@909: if $CODERAY_DEBUG && (@opened.empty? || @opened.last != kind) Chris@909: warn 'Malformed token stream: Trying to close a line (%p) ' \ Chris@909: 'that is not open. Open are: %p.' % [kind, @opened[1..-1]] Chris@909: end Chris@909: if @opened.pop Chris@909: @out << '' Chris@909: @last_opened = @opened.last if @last_opened Chris@909: end Chris@909: end Chris@909: Chris@909: end Chris@909: Chris@909: end Chris@909: end