RedCloth 3.0.4 → 4.0.0

data/lib/redcloth/formatters/base.rb

@@ -0,0 +1,50 @@
+module RedCloth::Formatters
+  module Base
+    
+    def pba(opts)
+      opts.delete(:style) if filter_styles
+      opts.delete(:class) if filter_classes
+      opts.delete(:id) if filter_ids
+
+      atts = ''
+      opts[:"text-align"] = opts.delete(:align)
+      opts[:style] += ';' if opts[:style] && (opts[:style][-1..-1] != ';')
+      [:float, :"text-align", :"vertical-align"].each do |a|
+        opts[:style] = "#{a}:#{opts[a]};#{opts[:style]}" if opts[a]
+      end
+      [:"padding-right", :"padding-left"].each do |a|
+        opts[:style] = "#{a}:#{opts[a]}em;#{opts[:style]}" if opts[a]
+      end
+      [:style, :class, :lang, :id, :colspan, :rowspan, :title, :start, :align].each do |a|
+        atts << " #{a}=\"#{ opts[a] }\"" if opts[a]
+      end
+      atts
+    end
+    
+    def ignore(opts)
+      opts[:text]
+    end
+    alias_method :notextile, :ignore
+    
+    def redcloth_version(opts)
+      p(:text => "#{opts[:prefix]}#{RedCloth::VERSION::STRING}")
+    end
+
+    def inline_redcloth_version(opts)
+      RedCloth::VERSION::STRING
+    end
+    
+    def method_missing(method, opts)
+      opts[:text] || ""
+    end
+    
+    def before_transform(text)
+      
+    end
+    
+    def after_transform(text)
+      
+    end
+
+  end
+end

data/lib/redcloth/formatters/html.rb

@@ -0,0 +1,342 @@
+module RedCloth::Formatters::HTML
+  include RedCloth::Formatters::Base
+  
+  # escapement for regular HTML (not in PRE tag)
+  def escape(text)
+    html_esc(text)
+  end
+
+  # escapement for HTML in a PRE tag
+  def escape_pre(text)
+    html_esc(text, :html_escape_preformatted)
+  end
+  
+  def after_transform(text)
+    text.chomp!
+  end
+    
+  [:h1, :h2, :h3, :h4, :h5, :h6, :p, :pre, :div].each do |m|
+    define_method(m) do |opts|
+      "<#{m}#{pba(opts)}>#{opts[:text]}</#{m}>\n"
+    end
+  end
+  
+  [:strong, :code, :em, :i, :b, :ins, :sup, :sub, :span, :cite].each do |m|
+    define_method(m) do |opts|
+      opts[:block] = true
+      "<#{m}#{pba(opts)}>#{opts[:text]}</#{m}>"
+    end
+  end
+  
+  def acronym(opts)
+    opts[:block] = true
+    "<acronym#{pba(opts)}>#{caps(:text => opts[:text])}</acronym>"
+  end
+  
+  def caps(opts)
+    if no_span_caps
+      opts[:text]
+    else
+      opts[:class] = 'caps'
+      span(opts)
+    end
+  end
+  
+  def del(opts)
+    opts[:block] = true
+    "<del#{pba(opts)}>#{opts[:text]}</del>"
+  end
+  
+  def del_phrase(opts)
+    " #{del(opts)}"
+  end
+  
+  [:ol, :ul].each do |m|
+    define_method("#{m}_open") do |opts|
+      opts[:block] = true
+      "#{"\n" if opts[:nest] > 1}#{"\t" * (opts[:nest] - 1)}<#{m}#{pba(opts)}>\n"
+    end
+    define_method("#{m}_close") do |opts|
+      "#{li_close}#{"\t" * (opts[:nest] - 1)}</#{m}>#{"\n" if opts[:nest] <= 1}"
+    end
+  end
+  
+  def li_open(opts)
+    "#{li_close unless opts.delete(:first)}#{"\t" * opts[:nest]}<li#{pba(opts)}>#{opts[:text]}"
+  end
+  
+  def li_close(opts=nil)
+    "</li>\n"
+  end
+  
+  def dl_open(opts)
+    opts[:block] = true
+    "<dl#{pba(opts)}>\n"
+  end
+  
+  def dl_close(opts=nil)
+    "</dl>\n"
+  end
+  
+  [:dt, :dd].each do |m|
+    define_method(m) do |opts|
+      "\t<#{m}#{pba(opts)}>#{opts[:text]}</#{m}>\n"
+    end
+  end
+  
+  def td(opts)
+    tdtype = opts[:th] ? 'th' : 'td'
+    "\t\t<#{tdtype}#{pba(opts)}>#{opts[:text]}</#{tdtype}>\n"
+  end
+  
+  def tr_open(opts)
+    "\t<tr#{pba(opts)}>\n"
+  end
+  
+  def tr_close(opts)
+    "\t</tr>\n"
+  end
+  
+  def table_open(opts)
+    "<table#{pba(opts)}>\n"
+  end
+  
+  def table_close(opts)
+    "</table>\n"
+  end
+  
+  def bc_open(opts)
+    opts[:block] = true
+    "<pre#{pba(opts)}>"
+  end
+  
+  def bc_close(opts)
+    "</pre>\n"
+  end
+  
+  def bq_open(opts)
+    opts[:block] = true
+    cite = opts[:cite] ? " cite=\"#{ opts[:cite] }\"" : ''
+    "<blockquote#{cite}#{pba(opts)}>\n"
+  end
+  
+  def bq_close(opts)
+    "</blockquote>\n"
+  end
+  
+  LINK_TEXT_WITH_TITLE_RE = /
+          ([^"]+?)         # $text
+          \s?
+          \(([^)]+?)\)     # $title
+          $
+      /x
+  def link(opts)
+    if opts[:name] =~ LINK_TEXT_WITH_TITLE_RE
+      md = LINK_TEXT_WITH_TITLE_RE.match(opts[:name])
+      opts[:name] = md[1]
+      opts[:title] = md[2]
+    end
+    "<a href=\"#{opts[:href].gsub(/&/, '&#38;')}\"#{pba(opts)}>#{opts[:name]}</a>"
+  end
+  
+  def image(opts)
+    opts.delete(:align)
+    opts[:alt] = opts[:title]
+    img = "<img src=\"#{urlesc opts[:src]}\"#{pba(opts)} alt=\"#{opts[:alt]}\" />"  
+    img = "<a href=\"#{urlesc opts[:href]}\">#{img}</a>" if opts[:href]
+    img
+  end
+  
+  def footno(opts)
+    opts[:id] ||= opts[:text]
+    %Q{<sup class="footnote"><a href=\"#fn#{opts[:id]}\">#{opts[:text]}</a></sup>}
+  end
+  
+  def fn(opts)
+    no = opts[:id]
+    opts[:id] = "fn#{no}"
+    opts[:class] = ["footnote", opts[:class]].compact.join(" ")
+    "<p#{pba(opts)}><sup>#{no}</sup> #{opts[:text]}</p>\n"
+  end
+  
+  def snip(opts)
+    "<pre#{pba(opts)}><code>#{opts[:text]}</code></pre>\n"
+  end
+  
+  def quote1(opts)
+    "&#8216;#{opts[:text]}&#8217;"
+  end
+  
+  def quote2(opts)
+    "&#8220;#{opts[:text]}&#8221;"
+  end
+  
+  def ellipsis(opts)
+    "#{opts[:text]}&#8230;"
+  end
+  
+  def emdash(opts)
+    "&#8212;"
+  end
+  
+  def endash(opts)
+    " &#8211; "
+  end
+  
+  def arrow(opts)
+    "&#8594;"
+  end
+  
+  def dim(opts)
+    space = opts[:space] ? " " : ''
+    "#{opts[:text]}#{space}&#215;#{space}"
+  end
+  
+  def trademark(opts)
+    "&#8482;"
+  end
+  
+  def registered(opts)
+    "&#174;"
+  end
+  
+  def copyright(opts)
+    "&#169;"
+  end
+  
+  def entity(opts)
+    "&#{opts[:text]};"
+  end
+  
+  def urlesc(txt)
+    txt.gsub(/&/, '&amp;')
+  end
+  
+  def amp(opts)
+    "&amp;"
+  end
+  
+  def gt(opts)
+    "&gt;"
+  end
+  
+  def lt(opts)
+    "&lt;"
+  end
+  
+  def br(opts)
+    if hard_breaks == false
+      "\n"
+    else
+      "<br />\n"
+    end
+  end
+  
+  def quot(opts)
+    "&quot;"
+  end
+  
+  def squot(opts)
+    "&#8217;"
+  end
+  
+  def html(opts)
+    "#{opts[:text]}\n"
+  end
+  
+  def html_block(opts)
+    inline_html(:text => "#{opts[:indent_before_start]}#{opts[:start_tag]}#{opts[:indent_after_start]}") + 
+    "#{opts[:text]}" +
+    inline_html(:text => "#{opts[:indent_before_end]}#{opts[:end_tag]}#{opts[:indent_after_end]}")
+  end
+  
+  def notextile(opts)
+    if filter_html
+      html_esc(opts[:text], :html_escape_preformatted)
+    else
+      opts[:text]
+    end
+  end
+  
+  def inline_html(opts)
+    if filter_html
+      html_esc(opts[:text], :html_escape_preformatted)
+    else
+      "#{opts[:text]}" # nil-safe
+    end
+  end
+  
+  def ignored_line(opts)
+    opts[:text] + "\n"
+  end
+  
+  def before_transform(text)
+    clean_html(text) if sanitize_html
+  end
+  
+  # HTML cleansing stuff
+  BASIC_TAGS = {
+      'a' => ['href', 'title'],
+      'img' => ['src', 'alt', 'title'],
+      'br' => [],
+      'i' => nil,
+      'u' => nil, 
+      'b' => nil,
+      'pre' => nil,
+      'kbd' => nil,
+      'code' => ['lang'],
+      'cite' => nil,
+      'strong' => nil,
+      'em' => nil,
+      'ins' => nil,
+      'sup' => nil,
+      'sub' => nil,
+      'del' => nil,
+      'table' => nil,
+      'tr' => nil,
+      'td' => ['colspan', 'rowspan'],
+      'th' => nil,
+      'ol' => ['start'],
+      'ul' => nil,
+      'li' => nil,
+      'p' => nil,
+      'h1' => nil,
+      'h2' => nil,
+      'h3' => nil,
+      'h4' => nil,
+      'h5' => nil,
+      'h6' => nil, 
+      'blockquote' => ['cite'],
+      'notextile' => nil
+  }
+  
+  # Clean unauthorized tags.
+  def clean_html( text, allowed_tags = BASIC_TAGS )
+    text.gsub!( /<!\[CDATA\[/, '' )
+    text.gsub!( /<(\/*)([A-Za-z]\w*)([^>]*?)(\s?\/?)>/ ) do |m|
+      raw = $~
+      tag = raw[2].downcase
+      if allowed_tags.has_key? tag
+        pcs = [tag]
+        allowed_tags[tag].each do |prop|
+          ['"', "'", ''].each do |q|
+            q2 = ( q != '' ? q : '\s' )
+            if raw[3] =~ /#{prop}\s*=\s*#{q}([^#{q2}]+)#{q}/i
+              attrv = $1
+              next if (prop == 'src' or prop == 'href') and not attrv =~ %r{^(http|https|ftp):}
+              pcs << "#{prop}=\"#{attrv.gsub('"', '\\"')}\""
+              break
+            end
+          end
+        end if allowed_tags[tag]
+        "<#{raw[1]}#{pcs.join " "}#{raw[4]}>"
+      else # Unauthorized tag
+        if block_given?
+          yield m
+        else
+          ''
+        end
+      end
+    end
+  end
+end

data/lib/redcloth/formatters/latex.rb

@@ -0,0 +1,227 @@
+require 'yaml' 
+
+module RedCloth::Formatters::LATEX
+  include RedCloth::Formatters::Base
+  
+  ENTITIES = YAML::load(File.read(File.dirname(__FILE__)+'/latex_entities.yml')) 
+  
+  def escape(text)
+    latex_esc(text)
+  end
+
+  def escape_pre(text)
+    text
+  end
+  
+  # headers
+  { :h1 => 'section*',
+    :h2 => 'subsection*',
+    :h3 => 'subsubsection*',
+    :h4 => 'textbf',
+    :h5 => 'textbf',
+    :h6 => 'textbf',
+  }.each do |m,tag| 
+    define_method(m) do |opts| 
+      "\\#{tag}{#{opts[:text]}}\n\n" 
+    end 
+  end 
+  
+  # commands 
+  { :strong => 'textbf',
+    :em => 'emph',
+    :i  => 'emph',
+    :b  => 'textbf',
+    :ins => 'underline',
+    :del => 'sout',
+    :acronym => 'MakeUppercase',
+    :caps => 'MakeUppercase',
+  }.each do |m,tag|
+    define_method(m) do |opts|
+      "\\#{tag}{#{opts[:text]}}"
+    end
+  end
+  
+  { :sup => '\ensuremath{^\textrm{#1}}',
+    :sub => '\ensuremath{_\textrm{#1}}',
+  }.each do |m, expr|
+    define_method(m) do |opts|
+      expr.sub('#1', opts[:text])
+    end
+  end
+  
+  # environments
+  { :pre  => 'verbatim',
+    :code => 'verbatim',
+    :cite => 'quote',
+    }.each do |m, env|
+    define_method(m) do |opts|
+      begin_chunk(env) + opts[:text] + end_chunk(env)
+    end
+  end
+  
+  # ignore (or find a good solution later)
+  [ :span,
+    :div,
+    ].each do |m|
+    define_method(m) do |opts|
+      opts[:text].to_s
+    end
+  end
+  
+  def del_phrase(opts)
+    " #{del(opts)}"
+  end
+  
+  { :ol => 'enumerate',
+    :ul => 'itemize',
+    }.each do |m, env|
+    define_method("#{m}_open") do |opts|
+      opts[:block] = true
+      "\\begin{#{env}}\n"
+    end
+    define_method("#{m}_close") do |opts|
+      "#{li_close}\\end{#{env}}\n\n"
+    end
+  end
+  
+  def li_open(opts)
+    "#{li_close unless opts.delete(:first)}\t\\item #{opts[:text]}"
+  end
+  
+  def li_close(opts=nil)
+    "\n"
+  end
+  
+  def p(opts)
+    opts[:text] + "\n\n"
+  end
+  
+  def td(opts)
+    "\t\t\t#{opts[:text]} &\n"
+  end
+  
+  def tr_open(opts)
+    "\t\t"
+  end
+  
+  def tr_close(opts)
+    "\t\t\\\\\n"
+  end
+  
+  # FIXME: we need to know the column count before opening tabular context.
+  def table_open(opts)
+    "\\begin{align*}\n"
+  end
+  
+  def table_close(opts)
+    "\t\\end{align*}\n"
+  end
+
+  def bc_open(opts)
+    opts[:block] = true
+    begin_chunk("verbatim") + "\n"
+  end
+
+  def bc_close(opts)
+    end_chunk("verbatim") + "\n"
+  end
+
+  def bq_open(opts)
+    opts[:block] = true
+    "\\begin{quotation}\n"
+  end
+
+  def bq_close(opts)
+    "\\end{quotation}\n\n"
+  end
+  
+  def link(opts)
+    "\\href{#{opts[:href]}}{#{opts[:name]}}"
+  end
+  
+  # FIXME: use includegraphics with security verification
+  def image(opts)
+    ""
+  end
+  
+  def footno(opts)
+    # TODO: insert a placeholder until we know the footnote content.
+    # For this to work, we need some kind of post-processing...
+    "\\footnotemark[#{opts[:text]}]"
+  end
+  
+  def fn(opts)
+    "\\footnotetext[#{opts[:id]}]{#{opts[:text]}}"
+  end
+  
+  def snip(opts)
+    "\\begin{verbatim}#{opts[:text]}\\end{verbatim}"
+  end
+  
+  def quote1(opts)
+    "`#{opts[:text]}'"
+  end
+  
+  def quote2(opts)
+    "``#{opts[:text]}\""
+  end
+  
+  def ellipsis(opts)
+    "#{opts[:text]}\\ldots{}"
+  end
+  
+  def emdash(opts)
+    "---"
+  end
+  
+  def endash(opts)
+    "--"
+  end
+  
+  def arrow(opts)
+    "\\rightarrow{}"
+  end
+  
+  def trademark(opts)
+    "\\texttrademark{}"
+  end
+  
+  def registered(opts)
+    "\\textregistered{}"
+  end
+  
+  def copyright(opts)
+    "\\copyright{}"
+  end
+  
+  # TODO: what do we do with (unknown) unicode entities ? 
+  #
+  def entity(opts)
+    text = opts[:text][0..0] == '#' ? opts[:text][1..-1] : opts[:text] 
+    ENTITIES[text]
+  end
+  
+  def dim(opts)
+    space = opts[:space] ? " " : ''
+    "#{opts[:text]}#{space}\\texttimes{}#{space}"
+  end
+  
+  private
+  
+  def begin_chunk(type)
+    chunk_counter[type] += 1
+    return "\\begin{#{type}}" if 1 == chunk_counter[type]
+    ''
+  end
+  
+  def end_chunk(type)
+    chunk_counter[type] -= 1
+    raise RuntimeError, "Bad latex #{type} nesting detected" if chunk_counter[type] < 0 # This should never need to happen
+    return "\\end{#{type}}" if 0 == chunk_counter[type]
+    ''
+  end
+  
+  def chunk_counter
+    @chunk_counter ||= Hash.new 0
+  end
+end