1  require 'rexml/formatters/default'
   2 
   3  module REXML
   4    module Formatters
   5      # Pretty-prints an XML document.  This destroys whitespace in text nodes
   6      # and will insert carriage returns and indentations.
   7      #
   8      # TODO: Add an option to print attributes on new lines
   9      class Pretty < Default
  10 
  11        # If compact is set to true, then the formatter will attempt to use as
  12        # little space as possible
  13        attr_accessor :compact
  14        # The width of a page.  Used for formatting text
  15        attr_accessor :width
  16 
  17        # Create a new pretty printer.
  18        #
  19        # output::
  20        #   An object implementing '<<(String)', to which the output will be written.
  21        # indentation::
  22        #   An integer greater than 0.  The indentation of each level will be
  23        #   this number of spaces.  If this is < 1, the behavior of this object
  24        #   is undefined.  Defaults to 2.
  25        # ie_hack::
  26        #   If true, the printer will insert whitespace before closing empty
  27        #   tags, thereby allowing Internet Explorer's feeble XML parser to
  28        #   function. Defaults to false.
  29        def initialize( indentation=2, ie_hack=false )
  30          @indentation = indentation
  31          @level = 0
  32          @ie_hack = ie_hack
  33          @width = 80
  34        end
  35 
  36        protected
  37        def write_element(node, output)
  38          output << ' '*@level
  39          output << "<#{node.expanded_name}"
  40 
  41          node.attributes.each_attribute do |attr|
  42            output << " "
  43            attr.write( output )
  44          end unless node.attributes.empty?
  45 
  46          if node.children.empty?
  47            if @ie_hack
  48              output << " "
  49            end
  50            output << "/" 
  51          else
  52            output << ">"
  53            # If compact and all children are text, and if the formatted output
  54            # is less than the specified width, then try to print everything on
  55            # one line
  56            skip = false
  57            if compact
  58              if node.children.inject(true) {|s,c| s & c.kind_of?(Text)}
  59                string = ""
  60                old_level = @level
  61                @level = 0
  62                node.children.each { |child| write( child, string ) }
  63                @level = old_level
  64                if string.length < @width
  65                  output << string
  66                  skip = true
  67                end
  68              end
  69            end
  70            unless skip
  71              output << "\n"
  72              @level += @indentation
  73              node.children.each { |child|
  74                next if child.kind_of?(Text) and child.to_s.strip.length == 0
  75                write( child, output )
  76                output << "\n"
  77              }
  78              @level -= @indentation
  79              output << ' '*@level
  80            end
  81            output << "</#{node.expanded_name}"
  82          end
  83          output << ">"
  84        end
  85 
  86        def write_text( node, output )
  87          s = node.to_s()
  88          s.gsub!(/\s/,' ')
  89          s.squeeze!(" ")
  90          s = wrap(s, 80-@level)
  91          s = indent_text(s, @level, " ", true)
  92          output << (' '*@level + s)
  93        end
  94 
  95        def write_comment( node, output)
  96          output << ' ' * @level
  97          super
  98        end
  99 
 100        def write_cdata( node, output)
 101          output << ' ' * @level
 102          super
 103        end
 104 
 105        def write_document( node, output )
 106          # Ok, this is a bit odd.  All XML documents have an XML declaration,
 107          # but it may not write itself if the user didn't specifically add it,
 108          # either through the API or in the input document.  If it doesn't write
 109          # itself, then we don't need a carriage return... which makes this
 110          # logic more complex.
 111          node.children.each { |child|
 112            next if child == node.children[-1] and child.instance_of?(Text)
 113            unless child == node.children[0] or child.instance_of?(Text) or
 114              (child == node.children[1] and !node.children[0].writethis)
 115              output << "\n"
 116            end
 117            write( child, output )
 118          }
 119        end
 120 
 121        private
 122        def indent_text(string, level=1, style="\t", indentfirstline=true)
 123          return string if level < 0
 124          string.gsub(/\n/, "\n#{style*level}")
 125        end
 126 
 127        def wrap(string, width)
 128          # Recursively wrap string at width.
 129          return string if string.length <= width
 130          place = string.rindex(' ', width) # Position in string with last ' ' before cutoff
 131          return string if place.nil?
 132          return string[0,place] + "\n" + wrap(string[place+1..-1], width)
 133        end
 134 
 135      end
 136    end
 137  end
 138