Method: HTML5::HTMLTokenizer#attribute_name_state

Defined in:

lib/html5/tokenizer.rb

#attribute_name_state ⇒ Object

# File 'lib/html5/tokenizer.rb', line 444

def attribute_name_state
  data = @stream.char
  leavingThisState = true
  emitToken = false
  if data == "="
    @state = :before_attribute_value_state
  elsif data == :EOF
    @token_queue << {:type => :ParseError, :data => "eof-in-attribute-name"}
    @state = :data_state
    emitToken = true
  elsif ASCII_LETTERS.include? data
    @current_token[:data][-1][0] += data + @stream.chars_until(ASCII_LETTERS, true)
    leavingThisState = false
  elsif data == ">"
    # XXX If we emit here the attributes are converted to a dict
    # without being checked and when the code below runs we error
    # because data is a dict not a list
    emitToken = true
  elsif SPACE_CHARACTERS.include? data
    @state = :after_attribute_name_state
  elsif data == "/"
    process_solidus_in_tag
    @state = :before_attribute_name_state
  else
    @current_token[:data][-1][0] += data
    leavingThisState = false
  end

  if leavingThisState
    # Attributes are not dropped at this stage. That happens when the
    # start tag token is emitted so values can still be safely appended
    # to attributes, but we do want to report the parse error in time.
    if @lowercase_attr_name
        @current_token[:data][-1][0] = @current_token[:data].last.first.downcase
    end
    @current_token[:data][0...-1].each {|name,value|
      if @current_token[:data].last.first == name
        @token_queue << {:type => :ParseError, :data => "duplicate-attribute"}
        break # don't report an error more than once
      end
    }
    # XXX Fix for above XXX
    emit_current_token if emitToken
  end
  return true
end