A nanoc filter for stripping blank lines and extraneous whitespace.

compact.rb

module Nanoc::Filters

    class Compact < Nanoc::Filter

        identifier :compact

        def run(content, params={})
            count_pre = 0
            content.split("\n").each do |line|
                line << "\n"
            
                line.match('<pre>')  { |m| count_pre +=1 }
                line.match('</pre>') { |m| count_pre -=1 }
            
                line.match('<code>')  { |m| count_pre +=1 }
                line.match('</code>') { |m| count_pre -=1 }
            
                line.match('<script')  { |m| count_pre +=1 }
                line.match('</script>') { |m| count_pre -=1 }
            
                # if we're inside any number of <pre> blocks, skip the reduction            
                if 0 == count_pre
                    line.gsub!(/[\r\n]/,' ')    # turn line endings into spaces
                    line.gsub!(/\s+/,' ')       # reduce serial whitespace to 1 character
                    line.gsub!(/>\s+</,'><')    # zap whitespace between tags
                    line.gsub!(/>\s+$/,'>')     # zap whitespace after tags
                    line.gsub!(/^\s+</,'<')     # zap whitespace before tags
                    line.gsub!(/^\s*$/,'')      # zap empty lines

                else
                    line.gsub!(/^\s+<pre>/  , '<pre>')
                    line.gsub!(/^\s+<code>/ , '<code>')
                    line.gsub!(/^\s+<script/, '<script')
                end
            end.join('')
        end

    end

end