Fix Unicode bug

Fix Diego Restrepo's bug (see Rev 184).
Update to latest HTML5lib.

vendor/plugins/HTML5lib/lib/html5/cli.rb

@@ -0,0 +1,231 @@
+$:.unshift File.dirname(__FILE__), 'lib'
+require 'html5'
+require 'core_ext/string'
+require 'ostruct'
+require 'optparse'
+
+module HTML5::CLI
+
+  def self.parse_opts argv
+    options = OpenStruct.new
+    options.profile        = false
+    options.time           = false
+    options.output         = :html
+    options.treebuilder    = 'simpletree'
+    options.error          = false
+    options.encoding       = false
+    options.parsemethod    = :parse
+    options.serializer     = {
+      :encoding            => 'utf-8',
+      :omit_optional_tags  => false,
+      :inject_meta_charset => false
+    }
+
+    opts = OptionParser.new do |opts|
+      opts.separator ""
+      opts.separator "Parse Options:"
+
+      opts.on("-b", "--treebuilder NAME") do |treebuilder|
+        options.treebuilder = treebuilder
+      end
+
+      opts.on("-f", "--fragment CONTAINER", "Parse as a fragment") do |container|
+        options.parsemethod = :parse_fragment
+        options.container = container if container
+      end
+
+      opts.separator ""
+      opts.separator "Filter Options:"
+
+      opts.on("--[no-]inject-meta-charset", "inject <meta charset>") do |inject|
+        options.serializer[:inject_meta_charset] = inject
+      end
+
+      opts.on("--[no-]strip-whitespace", "strip unnecessary whitespace") do |strip|
+        options.serializer[:strip_whitespace] = strip
+      end
+
+      opts.on("--[no-]sanitize", "escape unsafe tags") do |sanitize|
+        options.serializer[:sanitize] = sanitize
+      end
+
+      opts.separator ""
+      opts.separator "Output Options:"
+
+      opts.on("--tree", "output as debug tree") do |tree|
+        options.output = :tree
+      end
+
+      opts.on("-x", "--xml", "output as xml") do |xml|
+        options.output = :xml
+        options.treebuilder = "rexml"
+      end
+
+      opts.on("--[no-]html", "Output as html") do |html|
+        options.output = (html ? :html : nil)
+      end
+
+      opts.on("--hilite", "Output as formatted highlighted code.") do |hilite|
+        options.output = :hilite
+      end
+
+      opts.on("-e", "--error", "Print a list of parse errors") do |error|
+        options.error = error
+      end
+
+      opts.separator ""
+      opts.separator "Serialization Options:"
+
+      opts.on("--[no-]omit-optional-tags", "Omit optional tags") do |omit|
+        options.serializer[:omit_optional_tags] = omit
+      end
+
+      opts.on("--[no-]quote-attr-values", "Quote attribute values") do |quote|
+        options.serializer[:quote_attr_values] = quote
+      end
+
+      opts.on("--[no-]use-best-quote-char", "Use best quote character") do |best|
+        options.serializer[:use_best_quote_char] = best
+      end
+
+      opts.on("--quote-char C", "Use specified quote character") do |c|
+        options.serializer[:quote_char] = c
+      end
+
+      opts.on("--[no-]minimize-boolean-attributes", "Minimize boolean attributes") do |min|
+        options.serializer[:minimize_boolean_attributes] = min
+      end
+
+      opts.on("--[no-]use-trailing-solidus", "Use trailing solidus") do |slash|
+        options.serializer[:use_trailing_solidus] = slash
+      end
+
+      opts.on("--[no-]escape-lt-in-attrs", "Escape less than signs in attribute values") do |lt|
+        options.serializer[:escape_lt_in_attrs] = lt
+      end
+
+      opts.on("--[no-]escape-rcdata", "Escape rcdata element values") do |rcdata|
+        options.serializer[:escape_rcdata] = rcdata
+      end
+
+      opts.separator ""
+      opts.separator "Other Options:"
+
+      opts.on("-p", "--[no-]profile", "Profile the run") do |profile|
+        options.profile = profile
+      end
+
+      opts.on("-t", "--[no-]time", "Time the run") do |time|
+        options.time = time
+      end
+
+      opts.on("-c", "--[no-]encoding", "Print character encoding used") do |encoding|
+        options.encoding = encoding
+      end
+
+      opts.on_tail("-h", "--help", "Show this message") do
+        puts opts
+        exit
+      end
+
+
+    end
+    opts.parse!(argv)
+    options
+  end
+
+  def self.open_input f
+    if f
+      begin
+        if f[0..6] == 'http://'
+          require 'open-uri'
+          f = URI.parse(f).open
+          encoding = f.charset
+        elsif f == '-'
+          f = $stdin
+        else
+          f = open(f)
+        end
+      rescue
+      end
+    else
+      $stderr.write("No filename provided. Use -h for help\n")
+      exit(1)
+    end
+    f
+  end
+
+  def self.parse(opts, args)
+    encoding = nil
+
+    f = open_input args.last
+
+    require 'html5/treebuilders'
+    treebuilder = HTML5::TreeBuilders[opts.treebuilder]
+
+    if opts.output == :xml
+      require 'html5/liberalxmlparser'
+      p = HTML5::XMLParser.new(:tree=>treebuilder)
+    else
+      require 'html5/html5parser'
+      p = HTML5::HTMLParser.new(:tree=>treebuilder)
+    end
+
+    if opts.parsemethod == :parse
+      args = [f, encoding]
+    else
+      args = [f, (opts.container || 'div'), encoding]
+    end
+
+    if opts.profile
+      require 'profiler'
+      Profiler__::start_profile
+      p.send(opts.parsemethod, *args)
+      Profiler__::stop_profile
+      Profiler__::print_profile($stderr)
+    elsif opts.time
+      require 'time' # TODO: switch to benchmark
+      t0 = Time.new
+      document = p.send(opts.parsemethod, *args)
+      t1 = Time.new
+      print_output(p, document, opts)
+      t2 = Time.new
+      puts "\n\nRun took: %fs (plus %fs to print the output)"%[t1-t0, t2-t1]
+    else
+      document = p.send(opts.parsemethod, *args)
+      print_output(p, document, opts)
+    end
+  end
+
+  def self.print_output(parser, document, opts)
+    puts "Encoding: #{parser.tokenizer.stream.char_encoding}" if opts.encoding
+
+    case opts.output
+    when :xml
+      print document
+    when :html
+      require 'html5/treewalkers'
+      tokens = HTML5::TreeWalkers[opts.treebuilder].new(document)
+      require 'html5/serializer'
+      puts HTML5::HTMLSerializer.serialize(tokens, opts.serializer)
+    when :hilite
+      print document.hilite
+    when :tree
+      document = [document] unless document.respond_to?(:each)
+      document.each {|fragment| puts parser.tree.testSerializer(fragment)}
+    end
+
+    if opts.error
+      errList=[]
+      for pos, errorcode, datavars in parser.errors
+        errList << "Line #{pos[0]} Col #{pos[1]} " + (HTML5::E[errorcode] || "Unknown error \"#{errorcode}\"") % datavars
+      end
+      $stdout.write("\nParse errors:\n" + errList.join("\n")+"\n")
+    end
+  end
+
+  def self.run
+    options = parse_opts ARGV
+    parse options, ARGV
+  end
+end

vendor/plugins/HTML5lib/lib/html5/constants.rb

@@ -908,7 +908,7 @@ module HTML5
       "eof-in-bogus-doctype" =>
          _("Unexpected end of file in bogus doctype."),
       "eof-in-innerhtml" =>
-         _("XXX innerHTML EOF"),
+         _("Unexpected EOF in inner html mode."),
       "unexpected-doctype" =>
          _("Unexpected DOCTYPE. Ignored."),
       "non-html-root" =>
@@ -1040,7 +1040,8 @@ module HTML5
          _("Unexpected end tag (%(name))" +
            ". Expected end of file."),
       "unexpected-end-table-in-caption" =>
-        _("Unexpected end table tag in caption. Generates implied end caption.")
+        _("Unexpected end table tag in caption. Generates implied end caption."),
+      "end-html-in-innerhtml" => _("Unexpected html end tag in inner html mode.")
   }
 
 end

vendor/plugins/HTML5lib/lib/html5/filters.rb

@@ -1 +0,0 @@
-require 'html5/filters/optionaltags'

vendor/plugins/HTML5lib/lib/html5/html5parser/after_body_phase.rb

@@ -25,7 +25,7 @@ module HTML5
 
     def endTagHtml(name)
       if @parser.inner_html
-        parse_error
+        parse_error "end-html-in-innerhtml"
       else
         # XXX: This may need to be done, not sure
         # Don't set last_phase to the current phase but to the inBody phase

vendor/plugins/HTML5lib/lib/html5/html5parser/in_body_phase.rb

@@ -51,34 +51,22 @@ module HTML5
       super(parser, tree)
 
       # for special handling of whitespace in <pre>
-      if $-w
-        $-w = false
-        class << self; alias processSpaceCharactersNonPre processSpaceCharacters; end
-        $-w = true
-      else
-        class << self; alias processSpaceCharactersNonPre processSpaceCharacters; end
+      silence do
+        class << self
+          alias processSpaceCharactersNonPre processSpaceCharacters
+        end
       end
     end
 
     def processSpaceCharactersDropNewline(data)
       # #Sometimes (start of <pre> blocks) we want to drop leading newlines
 
-      if $-w
-        $-w = false
-        class << self
-          silence do
-            alias processSpaceCharacters processSpaceCharactersNonPre
-          end
-        end
-        $-w = true
-      else
-        class << self
-          silence do
-            alias processSpaceCharacters processSpaceCharactersNonPre
-          end
+      class << self
+        silence do
+          alias processSpaceCharacters processSpaceCharactersNonPre
         end
       end
-
+      
       if (data.length > 0 and data[0] == ?\n && 
         %w[pre textarea].include?(@tree.open_elements.last.name) && !@tree.open_elements.last.hasContent)
         data = data[1..-1]
@@ -376,16 +364,6 @@ module HTML5
     end
 
     def endTagBlock(name)
-      #Put us back in the right whitespace handling mode
-      if name == 'pre'
-        class << self; 
-          silence do
-            alias processSpaceCharacters processSpaceCharactersNonPre;
-          end
-        end
-      end
-      
-
       @tree.generateImpliedEndTags if in_scope?(name)
 
       unless @tree.open_elements.last.name == name

vendor/plugins/HTML5lib/lib/html5/html5parser/phase.rb

@@ -144,7 +144,7 @@ module HTML5
 
     def remove_open_elements_until(name=nil)
       finished = false
-      until finished
+      until finished || @tree.open_elements.length == 0
         element = @tree.open_elements.pop
         finished = name.nil? ? yield(element) : element.name == name
       end