premailer 1.6.2 → 1.7.0

data/README.rdoc

@@ -15,10 +15,13 @@ script is my solution.
 * A plain text version is created
   Optional
 
+=== Premailer 2.0 is coming
+
+I'm looking for input on a version 2.0 update to Premailer.  PLease visit the {Premailer 2.0 Planning Page}[https://github.com/alexdunae/premailer/wiki/Premailer-2.0-Planning] and give me your feedback.
 
 === Installation
 
-Download the Premailer gem from GemCutter.
+Download the Premailer gem from RubyGems.
 
     sudo gem install premailer
 
@@ -42,7 +45,7 @@ Download the Premailer gem from GemCutter.
 
 === Contributions
 
-Contributions are most welcome.  Premailer was rotting away in a private SVN repository for too long and could use some TLC.  Pull and patch to your heart's content.
+Contributions are most welcome.  Premailer was rotting away in a private SVN repository for too long and could use some TLC.  Fork and patch to your heart's content.  Please don't increment the version numbers, though.
 
 A few areas that are particularly in need of love:
 * Testing suite
@@ -54,13 +57,13 @@ A few areas that are particularly in need of love:
 
 === Credits and code
 
-Thanks to {all the wonderful contributors}[http://github.com/alexdunae/premailer/contributors] for their updates.
+Thanks to {all the wonderful contributors}[https://github.com/alexdunae/premailer/contributors] for their updates.
 
 Thanks to {Greenhood + Company}[http://www.greenhood.com/] for sponsoring some of the 1.5.6 updates, 
 and to {Campaign Monitor}[http://www.campaignmonitor.com] for supporting the web interface.
 
-The web interface can be found at http://premailer.dialect.ca .
+The web interface can be found at {premailer.dialect.ca}[http://premailer.dialect.ca].
 
-The source code can be found at http://github.com/alexdunae/premailer .
+The source code can be found on {GitHub}[https://github.com/alexdunae/premailer].
 
-Written by Alex Dunae (dunae.ca, e-mail 'code' at the same domain), 2008-2010.
+Copyright by Alex Dunae (dunae.ca, e-mail 'code' at the same domain), 2007-2011.  See LICENSE.rdoc for license details.

data/bin/premailer

@@ -31,7 +31,7 @@ opts = OptionParser.new do |opts|
     mode = v
   end
 
-  opts.on("-b", "--base-url", String, "Base URL, useful for local files") do |v|
+  opts.on("-b", "--base-url STRING", String, "Base URL, useful for local files") do |v|
     options[:base_url] = v
   end
 

data/lib/premailer.rb

@@ -1,7 +1,11 @@
 require 'yaml'
 require 'open-uri'
+require 'digest/md5'
 require 'cgi'
-require 'hpricot'
 require 'css_parser'
+require File.expand_path(File.dirname(__FILE__) + '/premailer/adapter')
+require File.expand_path(File.dirname(__FILE__) + '/premailer/adapter/hpricot')
+require File.expand_path(File.dirname(__FILE__) + '/premailer/adapter/nokogiri')
+
 require File.expand_path(File.dirname(__FILE__) + '/premailer/html_to_plain_text')
 require File.expand_path(File.dirname(__FILE__) + '/premailer/premailer')

data/lib/premailer/adapter.rb

@@ -0,0 +1,49 @@
+# = HTTPI::Adapter
+#
+# Manages the adapter classes. Currently supports:
+#
+# * nokogiri
+# * hpricot
+module Adapter
+    DEFAULT = :hpricot
+
+    # Returns the adapter to use. Defaults to <tt>Adapter::</tt>.
+    def self.use
+      @use ||= DEFAULT
+    end
+
+    # Sets the +adapter+ to use. Raises an +ArgumentError+ unless the +adapter+ exists.
+    def self.use=(adapter)
+      validate_adapter! adapter
+      @use = adapter
+    end
+
+    # Returns a memoized +Hash+ of adapters.
+    def self.adapters
+      @adapters ||= {
+        :nokogiri => { :class => Nokogiri, :require => "nokogiri" },
+        :hpricot  => { :class => Hpricot,  :require => "hpricot" },
+      }
+    end
+
+    # Returns an +adapter+. Raises an +ArgumentError+ unless the +adapter+ exists.
+    def self.find(adapter)
+      validate_adapter! adapter
+      load_adapter adapter
+    end
+
+  private
+
+    # Raises an +ArgumentError+ unless the +adapter+ exists.
+    def self.validate_adapter!(adapter)
+      raise ArgumentError, "Invalid adapter: #{adapter}" unless adapters[adapter]
+    end
+
+    # Tries to load and return the given +adapter+ name and class and falls back to the +FALLBACK+ adapter.
+    def self.load_adapter(adapter)
+      require adapters[adapter][:require]
+      [adapter, adapters[adapter][:class]]
+    rescue LoadError
+      puts "tried to use the #{adapter} adapter, but was unable to find the library in the LOAD_PATH."
+    end
+end

data/lib/premailer/adapter/hpricot.rb

@@ -0,0 +1,180 @@
+
+module Adapter
+	module Hpricot
+	
+	  # Merge CSS into the HTML document.
+  #
+  # Returns a string.
+  def to_inline_css
+    doc = @processed_doc
+    @unmergable_rules = CssParser::Parser.new
+    
+    # Give all styles already in style attributes a specificity of 1000 
+    # per http://www.w3.org/TR/CSS21/cascade.html#specificity
+    doc.search("*[@style]").each do |el| 
+      el['style'] = '[SPEC=1000[' + el.attributes['style'] + ']]'
+    end
+
+    # Iterate through the rules and merge them into the HTML
+    @css_parser.each_selector(:all) do |selector, declaration, specificity|
+      # Save un-mergable rules separately
+      selector.gsub!(/:link([\s]*)+/i) {|m| $1 }
+
+      # Convert element names to lower case
+      selector.gsub!(/([\s]|^)([\w]+)/) {|m| $1.to_s + $2.to_s.downcase }
+      
+      if selector =~ Premailer::RE_UNMERGABLE_SELECTORS
+        @unmergable_rules.add_rule_set!(CssParser::RuleSet.new(selector, declaration)) unless @options[:preserve_styles]
+      else
+        begin
+          # Change single ID CSS selectors into xpath so that we can match more 
+          # than one element.  Added to work around dodgy generated code.
+          selector.gsub!(/\A\#([\w_\-]+)\Z/, '*[@id=\1]')
+
+          doc.search(selector).each do |el|
+            if el.elem? and (el.name != 'head' and el.parent.name != 'head')
+              # Add a style attribute or append to the existing one  
+              block = "[SPEC=#{specificity}[#{declaration}]]"
+              el['style'] = (el.attributes['style'].to_s ||= '') + ' ' + block
+            end
+          end
+        rescue Hpricot::Error, RuntimeError, ArgumentError
+          $stderr.puts "CSS syntax error with selector: #{selector}" if @options[:verbose]
+          next
+        end
+      end
+    end
+
+    # Read STYLE attributes and perform folding
+    doc.search("*[@style]").each do |el|
+      style = el.attributes['style'].to_s
+      
+      declarations = []
+
+      style.scan(/\[SPEC\=([\d]+)\[(.[^\]\]]*)\]\]/).each do |declaration|
+        rs = CssParser::RuleSet.new(nil, declaration[1].to_s, declaration[0].to_i)
+        declarations << rs
+      end
+
+      # Perform style folding
+      merged = CssParser.merge(declarations)
+      merged.expand_shorthand!
+           
+      # Duplicate CSS attributes as HTML attributes
+      if Premailer::RELATED_ATTRIBUTES.has_key?(el.name)       
+        Premailer::RELATED_ATTRIBUTES[el.name].each do |css_att, html_att|
+          el[html_att] = merged[css_att].gsub(/;$/, '').strip if el[html_att].nil? and not merged[css_att].empty?
+        end
+      end
+      
+      merged.create_dimensions_shorthand!
+
+      # write the inline STYLE attribute
+      el['style'] = Premailer.escape_string(merged.declarations_to_s)
+    end
+
+    doc = write_unmergable_css_rules(doc, @unmergable_rules)
+
+    if @options[:remove_classes] or @options[:remove_comments]
+      doc.search('*').each do |el|
+        if el.comment? and @options[:remove_comments]
+          lst = el.parent.children
+          el.parent = nil
+          lst.delete(el)
+        elsif el.elem?
+          el.remove_attribute('class') if @options[:remove_classes]
+        end
+      end
+    end
+
+    if @options[:remove_ids]
+      # find all anchor's targets and hash them
+      targets = []
+      doc.search("a[@href^='#']").each do |el|
+        target = el.get_attribute('href')[1..-1]
+        targets << target
+        el.set_attribute('href', "#" + Digest::MD5.hexdigest(target))
+      end
+      # hash ids that are links target, delete others
+      doc.search("*[@id]").each do |el|
+        id = el.get_attribute('id')
+        if targets.include?(id)
+          el.set_attribute('id', Digest::MD5.hexdigest(id))
+        else
+          el.remove_attribute('id')
+        end
+      end
+    end
+
+    @processed_doc = doc
+
+    @processed_doc.to_original_html
+  end
+	
+	# Create a <tt>style</tt> element with un-mergable rules (e.g. <tt>:hover</tt>) 
+  # and write it into the <tt>body</tt>.
+  #
+  # <tt>doc</tt> is an Hpricot document and <tt>unmergable_css_rules</tt> is a Css::RuleSet.
+  #
+  # Returns an Hpricot document.
+  def write_unmergable_css_rules(doc, unmergable_rules) # :nodoc:
+    if head = doc.search('head')
+      styles = ''
+      unmergable_rules.each_selector(:all, :force_important => true) do |selector, declarations, specificity|
+        styles += "#{selector} { #{declarations} }\n"
+      end    
+
+      unless styles.empty?
+        style_tag = "\n<style type=\"text/css\">\n#{styles}</style>\n"
+        head.html.empty? ? head.inner_html(style_tag) : head.append(style_tag)
+      end
+    else
+      $stderr.puts "Unable to write unmergable CSS rules: no <head> was found" if @options[:verbose]
+    end
+    doc
+  end
+
+	
+    # Converts the HTML document to a format suitable for plain-text e-mail.
+    #
+		# If present, uses the <body> element as its base; otherwise uses the whole document.
+		#
+		# Returns a string.
+		def to_plain_text
+			html_src = ''
+			begin
+				html_src = @doc.search("body").inner_html
+			rescue; end
+
+			html_src = @doc.to_html unless html_src and not html_src.empty?
+			convert_to_text(html_src, @options[:line_length], @html_encoding)
+		end	
+	
+	
+	  # Returns the original HTML as a string.
+		def to_s
+			@doc.to_original_html
+		end
+
+	  # Load the HTML file and convert it into an Hpricot document.
+		#
+		# Returns an Hpricot document.
+		def load_html(input) # :nodoc:
+			thing = nil
+			
+			# TODO: duplicate options
+			if @options[:with_html_string] or @options[:inline] or input.respond_to?(:read)
+				thing = input
+				elsif @is_local_file
+				@base_dir = File.dirname(input)
+				thing = File.open(input, 'r')
+				else
+				thing = open(input)
+			end
+
+			# TODO: deal with Hpricot seg faults on empty input
+			thing ? Hpricot(thing) : nil  
+		end
+		
+	end
+end

data/lib/premailer/adapter/nokogiri.rb

@@ -0,0 +1,199 @@
+module Adapter
+	module Nokogiri
+	
+  	# Merge CSS into the HTML document.
+    #
+    # Returns a string.
+    def to_inline_css
+      doc = @processed_doc
+      @unmergable_rules = CssParser::Parser.new
+    
+      # Give all styles already in style attributes a specificity of 1000 
+      # per http://www.w3.org/TR/CSS21/cascade.html#specificity
+      doc.search("*[@style]").each do |el| 
+        el['style'] = '[SPEC=1000[' + el.attributes['style'] + ']]'
+      end
+
+      # Iterate through the rules and merge them into the HTML
+      @css_parser.each_selector(:all) do |selector, declaration, specificity|
+        # Save un-mergable rules separately
+        selector.gsub!(/:link([\s]*)+/i) {|m| $1 }
+
+        # Convert element names to lower case
+        selector.gsub!(/([\s]|^)([\w]+)/) {|m| $1.to_s + $2.to_s.downcase }
+      
+        if selector =~ Premailer::RE_UNMERGABLE_SELECTORS
+          @unmergable_rules.add_rule_set!(CssParser::RuleSet.new(selector, declaration)) unless @options[:preserve_styles]
+        else
+          begin
+            # Change single ID CSS selectors into xpath so that we can match more 
+            # than one element.  Added to work around dodgy generated code.
+            selector.gsub!(/\A\#([\w_\-]+)\Z/, '*[@id=\1]')
+
+            doc.search(selector).each do |el|
+              if el.elem? and (el.name != 'head' and el.parent.name != 'head')
+                # Add a style attribute or append to the existing one  
+                block = "[SPEC=#{specificity}[#{declaration}]]"
+                el['style'] = (el.attributes['style'].to_s ||= '') + ' ' + block
+              end
+            end
+          rescue  ::Nokogiri::SyntaxError, RuntimeError, ArgumentError
+            $stderr.puts "CSS syntax error with selector: #{selector}" if @options[:verbose]
+            next
+          end
+        end
+      end
+
+      # Read STYLE attributes and perform folding
+      doc.search("*[@style]").each do |el|
+        style = el.attributes['style'].to_s
+      
+        declarations = []
+
+        style.scan(/\[SPEC\=([\d]+)\[(.[^\]\]]*)\]\]/).each do |declaration|
+          rs = CssParser::RuleSet.new(nil, declaration[1].to_s, declaration[0].to_i)
+          declarations << rs
+        end
+
+        # Perform style folding
+        merged = CssParser.merge(declarations)
+        merged.expand_shorthand!
+          
+        # Duplicate CSS attributes as HTML attributes
+        if Premailer::RELATED_ATTRIBUTES.has_key?(el.name)       
+          Premailer::RELATED_ATTRIBUTES[el.name].each do |css_att, html_att|
+            el[html_att] = merged[css_att].gsub(/;$/, '').strip if el[html_att].nil? and not merged[css_att].empty?
+          end
+        end
+      
+        merged.create_dimensions_shorthand!
+
+        # write the inline STYLE attribute
+        el['style'] = Premailer.escape_string(merged.declarations_to_s)
+      end
+
+      doc = write_unmergable_css_rules(doc, @unmergable_rules)
+
+      if @options[:remove_classes] or @options[:remove_comments]
+        doc.traverse do |el|
+          if el.comment? and @options[:remove_comments]
+            el.remove
+          elsif el.element?
+            el.remove_attribute('class') if @options[:remove_classes]
+          end
+        end  
+      end
+
+      if @options[:remove_ids]
+        # find all anchor's targets and hash them
+        targets = []
+        doc.search("a[@href^='#']").each do |el|
+          target = el.get_attribute('href')[1..-1]
+          targets << target
+          el.set_attribute('href', "#" + Digest::MD5.hexdigest(target))
+        end
+        # hash ids that are links target, delete others
+        doc.search("*[@id]").each do |el|
+          id = el.get_attribute('id')
+          if targets.include?(id)
+            el.set_attribute('id', Digest::MD5.hexdigest(id))
+          else
+            el.remove_attribute('id')
+          end
+        end
+      end
+
+      @processed_doc = doc
+  		if is_xhtml?
+  			@processed_doc.to_xhtml
+  		else
+      	@processed_doc.to_html
+    	end
+    end
+	
+  	# Create a <tt>style</tt> element with un-mergable rules (e.g. <tt>:hover</tt>) 
+    # and write it into the <tt>body</tt>.
+    #
+    # <tt>doc</tt> is an Nokogiri document and <tt>unmergable_css_rules</tt> is a Css::RuleSet.
+    #
+    # Returns an Nokogiri document.
+    def write_unmergable_css_rules(doc, unmergable_rules) # :nodoc:
+      if head = doc.at('head')
+        styles = ''
+        unmergable_rules.each_selector(:all, :force_important => true) do |selector, declarations, specificity|
+          styles += "#{selector} { #{declarations} }\n"
+        end    
+
+        unless styles.empty?
+          style_tag = "\n<style type=\"text/css\">\n#{styles}</style>\n"
+
+          head.add_child(style_tag)
+        end
+      else
+        $stderr.puts "Unable to write unmergable CSS rules: no <head> was found" if @options[:verbose]
+      end
+      doc
+    end
+
+	
+    # Converts the HTML document to a format suitable for plain-text e-mail.
+    #
+    # If present, uses the <body> element as its base; otherwise uses the whole document.
+    #
+    # Returns a string.
+    def to_plain_text
+      html_src = ''
+      begin
+        html_src = @doc.at("body").inner_html
+      rescue; end
+
+      html_src = @doc.to_html unless html_src and not html_src.empty?
+      convert_to_text(html_src, @options[:line_length], @html_encoding)
+    end
+	
+		# Returns the original HTML as a string.
+		def to_s
+			if is_xhtml?
+				@doc.to_xhtml
+			else
+				@doc.to_html
+			end
+		end
+		
+	  # Load the HTML file and convert it into an Nokogiri document.
+		#
+		# Returns an Nokogiri document.
+		def load_html(input) # :nodoc:
+			thing = nil
+			
+			# TODO: duplicate options
+			if @options[:with_html_string] or @options[:inline] or input.respond_to?(:read)
+				thing = input
+				elsif @is_local_file
+				@base_dir = File.dirname(input)
+				thing = File.open(input, 'r')
+				else
+				thing = open(input)
+			end
+
+		  if thing.respond_to?(:read)
+		    thing = thing.read
+		  end
+			
+			return nil unless thing
+			
+			doc = nil
+
+      # Default encoding is ASCII-8BIT (binary) per http://groups.google.com/group/nokogiri-talk/msg/0b81ef0dc180dc74
+      if thing.is_a?(String) and RUBY_VERSION =~ /1.9/ 
+	      thing = thing.force_encoding('ASCII-8BIT').encode!
+		    doc = ::Nokogiri::HTML(thing) {|c| c.noent.recover }	    
+		  else
+		    doc = ::Nokogiri::HTML(thing, nil, 'ASCII-8BIT') {|c| c.noent.recover }
+	    end
+			
+			return doc
+		end
+		
+	end
+end

data/lib/premailer/html_to_plain_text.rb

@@ -6,15 +6,10 @@ module HtmlToPlainText
 
   # Returns the text in UTF-8 format with all HTML tags removed
   #
-  # TODO:
-  #  - add support for DL, OL
+  # TODO: add support for DL, OL
   def convert_to_text(html, line_length = 65, from_charset = 'UTF-8')
-    #r = Text::Reform.new(:trim => true,
-    #                     :squeeze => false,
-    #                     :break => Text::Reform.break_wrap)
-
     txt = html
-    
+
     # decode HTML entities
     he = HTMLEntities.new
     txt = he.decode(txt)
@@ -24,7 +19,7 @@ module HtmlToPlainText
     txt.gsub!(/[\s]*<h([1-6]+)[^>]*>[\s]*(.*)[\s]*<\/h[1-6]+>/i) do |s|
       hlevel = $1.to_i
 
-      htext = $2      
+      htext = $2
       htext.gsub!(/<br[\s]*\/?>/i, "\n") # handle <br>s
       htext.gsub!(/<\/?[^>]*>/i, '') # strip tags
 
@@ -41,10 +36,13 @@ module HtmlToPlainText
         else     # H3-H6, dashes below
           htext = htext + "\n" + ('-' * hlength)
       end
-      
+
       "\n\n" + htext + "\n\n"
     end
 
+    # wrap spans
+    txt.gsub!(/(<\/span>)[\s]+(<span)/mi, '\1 \2')
+
     # links
     txt.gsub!(/<a.*href=\"([^\"]*)\"[^>]*>(.*)<\/a>/i) do |s|
       $2.strip + ' ( ' + $1.strip + ' )'
@@ -54,20 +52,19 @@ module HtmlToPlainText
     txt.gsub!(/[\s]*(<li[^>]*>)[\s]*/i, '* ')
     # list not followed by a newline
     txt.gsub!(/<\/li>[\s]*(?![\n])/i, "\n")
-    
+
     # paragraphs and line breaks
     txt.gsub!(/<\/p>/i, "\n\n")
     txt.gsub!(/<br[\/ ]*>/i, "\n")
-    
+
     # strip remaining tags
     txt.gsub!(/<\/?[^>]*>/, '')
 
-    # wrap text
-    #txt = r.format(('[' * line_length), txt)
-    
+    txt = word_wrap(txt, line_length)
+
     # remove linefeeds (\r\n and \r -> \n)
     txt.gsub!(/\r\n?/, "\n")
-    
+
     # strip extra spaces
     txt.gsub!(/\302\240+/, " ") # non-breaking spaces -> spaces
     txt.gsub!(/\n[ \t]+/, "\n") # space at start of lines
@@ -78,4 +75,11 @@ module HtmlToPlainText
 
     txt.strip
   end
+
+  # Taken from Rails' word_wrap helper (http://api.rubyonrails.org/classes/ActionView/Helpers/TextHelper.html#method-i-word_wrap)
+  def word_wrap(txt, line_length)
+    txt.split("\n").collect do |line|
+      line.length > line_length ? line.gsub(/(.{1,#{line_length}})(\s+|$)/, "\\1\n").strip : line
+    end * "\n"
+  end
 end

data/lib/premailer/premailer.rb

@@ -4,8 +4,8 @@
 #
 # Premailer processes HTML and CSS to improve e-mail deliverability.
 #
-# Premailer's main function is to render all CSS as inline <tt>style</tt> 
-# attributes. It also converts relative links to absolute links and checks 
+# Premailer's main function is to render all CSS as inline <tt>style</tt>
+# attributes. It also converts relative links to absolute links and checks
 # the 'safety' of CSS properties against a CSS support chart.
 #
 # = Example
@@ -33,17 +33,17 @@ class Premailer
   include HtmlToPlainText
   include CssParser
 
-  VERSION = '1.6.2'
+  VERSION = '1.7.0'
 
   CLIENT_SUPPORT_FILE = File.dirname(__FILE__) + '/../../misc/client_support.yaml'
 
   RE_UNMERGABLE_SELECTORS = /(\:(visited|active|hover|focus|after|before|selection|target|first\-(line|letter))|^\@)/i
-  
+
   # list of CSS attributes that can be rendered as HTML attributes
   #
   # TODO: too much repetition
   # TODO: background=""
-  RELATED_ATTRIBUTES = { 
+  RELATED_ATTRIBUTES = {
     'h1' => {'text-align' => 'align'},
     'h2' => {'text-align' => 'align'},
     'h3' => {'text-align' => 'align'},
@@ -63,18 +63,20 @@ class Premailer
 
   # URI of the HTML file used
   attr_reader   :html_file
-  
+
   # base URL used to resolve links
   attr_reader   :base_url
 
   # base directory used to resolve links for local files
   attr_reader   :base_dir
 
-  
-  # processed HTML document (Hpricot)
+  # unmergeable CSS rules to be preserved in the head (CssParser)
+  attr_reader   :unmergable_rules
+
+  # processed HTML document (Hpricot/Nokogiri)
   attr_reader   :processed_doc
-  
-  # source HTML document (Hpricot)
+
+  # source HTML document (Hpricot/Nokogiri)
   attr_reader   :doc
 
   module Warnings
@@ -89,8 +91,8 @@ class Premailer
 
   # Create a new Premailer object.
   #
-  # +html+ is the HTML data to process. It can be either an IO object, the URL of a 
-  # remote file, a local path or a raw HTML string.  If passing an HTML string you 
+  # +html+ is the HTML data to process. It can be either an IO object, the URL of a
+  # remote file, a local path or a raw HTML string.  If passing an HTML string you
   # must set the +:with_html_string+ option to +true+.
   #
   # ==== Options
@@ -100,24 +102,33 @@ class Premailer
   # [+base_url+] Used to calculate absolute URLs for local files.
   # [+css+] Manually specify a CSS stylesheet.
   # [+css_to_attributes+] Copy related CSS attributes into HTML attributes (e.g. +background-color+ to +bgcolor+)
+  # [+css_string+] Pass CSS as a string
+  # [+remove_ids+] Remove ID attributes whenever possible and convert IDs used as anchors to hashed to avoid collisions in webmail programs.  Default is +false+.
+  # [+remove_classes+] Remove class attributes. Default is +false+.
+  # [+remove_comments+] Remove html comments. Default is +false+.
   # [+preserve_styles+] Whether to preserve any <tt>link rel=stylesheet</tt> and <tt>style</tt> elements.  Default is +false+.
   # [+with_html_string+] Whether the +html+ param should be treated as a raw string.
   # [+verbose+] Whether to print errors and warnings to <tt>$stderr</tt>.  Default is +false+.
+  # [+adapter+] Which HTML parser to use, either <tt>:nokogiri</tt> or <tt>:hpricot</tt>.  Default is <tt>:hpricot</tt>.
   def initialize(html, options = {})
-    @options = {:warn_level => Warnings::SAFE, 
-                :line_length => 65, 
-                :link_query_string => nil, 
+    @options = {:warn_level => Warnings::SAFE,
+                :line_length => 65,
+                :link_query_string => nil,
                 :base_url => nil,
                 :remove_classes => false,
+                :remove_ids => false,
+                :remove_comments => false,
                 :css => [],
                 :css_to_attributes => true,
                 :with_html_string => false,
+                :css_string => nil,
                 :preserve_styles => false,
                 :verbose => false,
                 :debug => false,
-                :io_exceptions => false}.merge(options)
+                :io_exceptions => false,
+                :adapter => Adapter.use}.merge(options)
 
-    @html_file = html 
+    @html_file = html
     @is_local_file = @options[:with_html_string] || Premailer.local_data?(html)
 
     @css_files = @options[:css]
@@ -126,6 +137,7 @@ class Premailer
 
     @base_url = nil
     @base_dir = nil
+    @unmergable_rules = nil
 
     if @options[:base_url]
       @base_url = URI.parse(@options.delete(:base_url))
@@ -138,10 +150,14 @@ class Premailer
       :import => true,
       :io_exceptions => @options[:io_exceptions]
     })
-    
+
+    @adapter_name = @options[:adapter]
+    @adapter_name, @adapter_class = Adapter.find @adapter_name
+
+    self.class.send(:include, @adapter_class)
+
     @doc = load_html(@html_file)
-    # TODO
-    @html_charset = nil # @doc.encoding || nil
+
     @processed_doc = @doc
     @processed_doc = convert_inline_links(@processed_doc, @base_url) if @base_url
     if options[:link_query_string]
@@ -157,141 +173,8 @@ class Premailer
     @css_warnings = check_client_support if @css_warnings.empty?
     @css_warnings
   end
-  
-  # Returns the original HTML as a string.
-  def to_s
-    @doc.to_original_html
-  end
-
-  # Converts the HTML document to a format suitable for plain-text e-mail.
-  #
-  # Returns a string.
-  def to_plain_text
-    html_src = ''
-    begin
-      html_src = @doc.search("body").inner_html
-    rescue
-      html_src = @doc.to_html
-    end
-    convert_to_text(html_src, @options[:line_length], @html_charset)
-  end
-
-  # Merge CSS into the HTML document.
-  #
-  # Returns a string.
-  def to_inline_css
-    doc = @processed_doc
-    unmergable_rules = CssParser::Parser.new
-    
-    # Give all styles already in style attributes a specificity of 1000 
-    # per http://www.w3.org/TR/CSS21/cascade.html#specificity
-    doc.search("*[@style]").each do |el| 
-      el['style'] = '[SPEC=1000[' + el.attributes['style'] + ']]'
-    end
-
-    # Iterate through the rules and merge them into the HTML
-    @css_parser.each_selector(:all) do |selector, declaration, specificity|
-      # Save un-mergable rules separately
-      selector.gsub!(/:link([\s]*)+/i) {|m| $1 }
-
-      # Convert element names to lower case
-      selector.gsub!(/([\s]|^)([\w]+)/) {|m| $1.to_s + $2.to_s.downcase }
-      
-      if selector =~ RE_UNMERGABLE_SELECTORS
-        unmergable_rules.add_rule_set!(RuleSet.new(selector, declaration)) unless @options[:preserve_styles]
-      else
-        begin
-          doc.search(selector).each do |el|
-            if el.elem? and (el.name != 'head' and el.parent.name != 'head')
-              # Add a style attribute or append to the existing one  
-              block = "[SPEC=#{specificity}[#{declaration}]]"
-              el['style'] = (el.attributes['style'].to_s ||= '') + ' ' + block
-            end
-          end
-        rescue Hpricot::Error, RuntimeError, ArgumentError
-          $stderr.puts "CSS syntax error with selector: #{selector}" if @options[:verbose]
-          next
-        end
-      end
-    end
-
-    # Read STYLE attributes and perform folding
-    doc.search("*[@style]").each do |el|
-      style = el.attributes['style'].to_s
-      
-      declarations = []
-
-      style.scan(/\[SPEC\=([\d]+)\[(.[^\]\]]*)\]\]/).each do |declaration|
-        rs = RuleSet.new(nil, declaration[1].to_s, declaration[0].to_i)
-        declarations << rs
-      end
-
-      # Perform style folding
-      merged = CssParser.merge(declarations)
-      merged.expand_shorthand!
-      
-      #if @options[:prefer_cellpadding] and (el.name == 'td' or el.name == 'th') and el['cellpadding'].nil?
-      #  if cellpadding = equivalent_cellpadding(merged)
-      #    el['cellpadding'] = cellpadding
-      #    merged['padding-left'] = nil
-      #    merged['padding-right'] = nil
-      #    merged['padding-top'] = nil
-      #    merged['padding-bottom'] = nil
-      #  end
-      #end
-      
-      # Duplicate CSS attributes as HTML attributes
-      if RELATED_ATTRIBUTES.has_key?(el.name)       
-        RELATED_ATTRIBUTES[el.name].each do |css_att, html_att|
-          el[html_att] = merged[css_att].gsub(/;$/, '').strip if el[html_att].nil? and not merged[css_att].empty?
-        end
-      end
-      
-      merged.create_dimensions_shorthand!
-
-      # write the inline STYLE attribute
-      el['style'] = Premailer.escape_string(merged.declarations_to_s)
-    end
-
-    doc = write_unmergable_css_rules(doc, unmergable_rules)
-
-    doc.search('*').remove_class if @options[:remove_classes]  
-
-    @processed_doc = doc
-
-    @processed_doc.to_original_html
-  end
-
-  # Check for an XHTML doctype
-  def is_xhtml?
-    intro = @doc.to_s.strip.split("\n")[0..2].join(' ')
-    is_xhtml = (intro =~ /w3c\/\/[\s]*dtd[\s]+xhtml/i)
-    is_xhtml = is_xhtml ? true : false
-    $stderr.puts "Is XHTML? #{is_xhtml.inspect}\nChecked:\n#{intro}" if @options[:debug]
-    is_xhtml
-  end
-
-protected  
-  # Load the HTML file and convert it into an Hpricot document.
-  #
-  # Returns an Hpricot document.
-  def load_html(input) # :nodoc:
-    thing = nil
-
-    # TODO: duplicate options
-    if @options[:with_html_string] or @options[:inline] or input.respond_to?(:read)
-      thing = input
-    elsif @is_local_file
-      @base_dir = File.dirname(input)
-      thing = File.open(input, 'r')
-    else
-      thing = open(input)
-    end
-
-    # TODO: deal with Hpricot seg faults on empty input
-    thing ? Hpricot(thing) : nil  
-  end
 
+protected
   def load_css_from_local_file!(path)
     css_block = ''
     begin
@@ -300,11 +183,18 @@ protected
           css_block << line
         end
       end
-      @css_parser.add_block!(css_block, {:base_uri => @base_url, :base_dir => @base_dir})
+
+      load_css_from_string(css_block)
     rescue; end
   end
 
+  def load_css_from_string(css_string)
+    @css_parser.add_block!(css_string, {:base_uri => @base_url, :base_dir => @base_dir, :only_media_types => [:screen, :handheld]})
+  end
+
   def load_css_from_options! # :nodoc:
+    load_css_from_string(@options[:css_string]) if @options[:css_string]
+
     @css_files.each do |css_file|
       if Premailer.local_data?(css_file)
         load_css_from_local_file!(css_file)
@@ -314,7 +204,7 @@ protected
     end
   end
 
-  # Load CSS included in <tt>style</tt> and <tt>link</tt> tags from an HTML document.
+    # Load CSS included in <tt>style</tt> and <tt>link</tt> tags from an HTML document.
   def load_css_from_html! # :nodoc:
     if tags = @doc.search("link[@rel='stylesheet'], style")
       tags.each do |tag|
@@ -326,10 +216,10 @@ protected
             load_css_from_local_file!(link_uri)
           else
             $stderr.puts "Loading css from uri: " + link_uri if @options[:verbose]
-            @css_parser.load_uri!(link_uri)
+            @css_parser.load_uri!(link_uri, {:only_media_types => [:screen, :handheld]})
           end
 
-        elsif tag.to_s.strip =~ /^\<style/i      
+        elsif tag.to_s.strip =~ /^\<style/i
           @css_parser.add_block!(tag.inner_html, :base_uri => @base_url, :base_dir => @base_dir, :only_media_types => [:screen, :handheld])
         end
       end
@@ -337,6 +227,17 @@ protected
     end
   end
 
+
+
+# here be deprecated methods
+public
+  def local_uri?(uri) # :nodoc:
+    warn "[DEPRECATION] `local_uri?` is deprecated.  Please use `Premailer.local_data?` instead."
+    Premailer.local_data?(uri)
+  end
+
+# here be instance methods
+
   def media_type_ok?(media_types) # :nodoc:
     return true if media_types.nil? or media_types.empty?
     return media_types.split(/[\s]+|,/).any? { |media_type| media_type.strip =~ /screen|handheld|all/i }
@@ -344,35 +245,12 @@ protected
     return true
   end
 
-  # Create a <tt>style</tt> element with un-mergable rules (e.g. <tt>:hover</tt>) 
-  # and write it into the <tt>body</tt>.
-  #
-  # <tt>doc</tt> is an Hpricot document and <tt>unmergable_css_rules</tt> is a Css::RuleSet.
-  #
-  # Returns an Hpricot document.
-  def write_unmergable_css_rules(doc, unmergable_rules) # :nodoc:
-    if head = doc.search('head')
-      styles = ''
-      unmergable_rules.each_selector(:all, :force_important => true) do |selector, declarations, specificity|
-        styles += "#{selector} { #{declarations} }\n"
-      end    
-
-      unless styles.empty?
-        style_tag = "\n<style type=\"text/css\">\n#{styles}</style>\n"
-        head.html.empty? ? head.inner_html(style_tag) : head.append(style_tag)
-      end
-    else
-      $stderr.puts "Unable to write unmergable CSS rules: no <head> was found" if @options[:verbose]
-    end
-    doc
-  end
-
   def append_query_string(doc, qs)
     return doc if qs.nil?
 
     qs.to_s.gsub!(/^[\?]*/, '').strip!
     return doc if qs.empty?
-    
+
     begin
       current_host = @base_url.host
     rescue
@@ -380,10 +258,10 @@ protected
     end
 
     $stderr.puts "Attempting to append_query_string: #{qs}" if @options[:verbose]
-    
+
     doc.search('a').each do|el|
       href = el.attributes['href'].to_s.strip
-      next if href.nil? or href.empty?    
+      next if href.nil? or href.empty?
       next if href[0,1] == '#' # don't bother with anchors
 
       begin
@@ -391,7 +269,7 @@ protected
 
         if current_host and href.host != nil and href.host != current_host
           $stderr.puts "Skipping append_query_string for: #{href.to_s} because host is no good" if @options[:verbose]
-          next 
+          next
         end
 
         if href.scheme and href.scheme != 'http' and href.scheme != 'https'
@@ -404,7 +282,7 @@ protected
         else
           href.query = qs
         end
-    
+
         el['href'] = href.to_s
       rescue URI::Error => e
         $stderr.puts "Skipping append_query_string for: #{href.to_s} (#{e.message})" if @options[:verbose]
@@ -415,9 +293,18 @@ protected
     doc
   end
 
-  # Convert relative links to absolute links.
+    # Check for an XHTML doctype
+  def is_xhtml?
+    intro = @doc.to_s.strip.split("\n")[0..2].join(' ')
+    is_xhtml = (intro =~ /w3c\/\/[\s]*dtd[\s]+xhtml/i)
+    is_xhtml = is_xhtml ? true : false
+    $stderr.puts "Is XHTML? #{is_xhtml.inspect}\nChecked:\n#{intro}" if @options[:debug]
+    is_xhtml
+  end
+
+    # Convert relative links to absolute links.
   #
-  # Processes <tt>href</tt> <tt>src</tt> and <tt>background</tt> attributes 
+  # Processes <tt>href</tt> <tt>src</tt> and <tt>background</tt> attributes
   # as well as CSS <tt>url()</tt> declarations found in inline <tt>style</tt> attributes.
   #
   # <tt>doc</tt> is an Hpricot document and <tt>base_uri</tt> is either a string or a URI.
@@ -466,20 +353,11 @@ protected
     doc
   end
 
-# here be deprecated methods
-public
-
-  def local_uri?(uri) # :nodoc:
-    warn "[DEPRECATION] `local_uri?` is deprecated.  Please use `Premailer.local_data?` instead."
-    Premailer.local_data?(uri)
-  end
-
-# here be instance methods
 
   def self.escape_string(str) # :nodoc:
-    str.gsub(/"/, "'")
+    str.gsub(/"/ , "'")
   end
-  
+
   def self.resolve_link(path, base_path) # :nodoc:
     path.strip!
     resolved = nil
@@ -488,7 +366,7 @@ public
       return Premailer.canonicalize(resolved)
     elsif base_path.kind_of?(URI)
       resolved = base_path.merge(path)
-      return Premailer.canonicalize(resolved)    
+      return Premailer.canonicalize(resolved)
     elsif base_path.kind_of?(String) and base_path =~ /^(http[s]?|ftp):\/\//i
       resolved = URI.parse(base_path)
       resolved = resolved.merge(path)
@@ -510,7 +388,7 @@ public
     else
       return true
     end
-  end  
+  end
 
   # from http://www.ruby-forum.com/topic/140101
   def self.canonicalize(uri) # :nodoc:
@@ -525,54 +403,51 @@ public
       u.to_s
     end
 
-    # Check <tt>CLIENT_SUPPORT_FILE</tt> for any CSS warnings
-    def check_client_support # :nodoc:
-      @client_support = @client_support ||= YAML::load(File.open(CLIENT_SUPPORT_FILE))
-
-      warnings = []
-      properties = []
-    
-      # Get a list off CSS properties
-      @processed_doc.search("*[@style]").each do |el|
-        style_url = el.attributes['style'].to_s.gsub(/([\w\-]+)[\s]*\:/i) do |s|
-          properties.push($1)
-        end
-      end
+  # Check <tt>CLIENT_SUPPORT_FILE</tt> for any CSS warnings
+  def check_client_support # :nodoc:
+    @client_support = @client_support ||= YAML::load(File.open(CLIENT_SUPPORT_FILE))
 
-      properties.uniq!
+    warnings = []
+    properties = []
 
-      property_support = @client_support['css_properties']
-      properties.each do |prop|
-        if property_support.include?(prop) and
-            property_support[prop].include?('support') and
-            property_support[prop]['support'] >= @options[:warn_level]
-          warnings.push({:message => "#{prop} CSS property",
-              :level => WARN_LABEL[property_support[prop]['support']],
-              :clients => property_support[prop]['unsupported_in'].join(', ')})
-        end
+    # Get a list off CSS properties
+    @processed_doc.search("*[@style]").each do |el|
+      style_url = el.attributes['style'].to_s.gsub(/([\w\-]+)[\s]*\:/i) do |s|
+        properties.push($1)
       end
+    end
 
-      @client_support['attributes'].each do |attribute, data|
-        next unless data['support'] >= @options[:warn_level]
-        if @doc.search("*[@#{attribute}]").length > 0
-          warnings.push({:message => "#{attribute} HTML attribute",
-              :level => WARN_LABEL[property_support[prop]['support']],
-              :clients => property_support[prop]['unsupported_in'].join(', ')})
-        end
-      end
+    properties.uniq!
 
-      @client_support['elements'].each do |element, data|
-        next unless data['support'] >= @options[:warn_level]
-        if @doc.search("element").length > 0
-          warnings.push({:message => "#{element} HTML element",
-              :level => WARN_LABEL[property_support[prop]['support']],
-              :clients => property_support[prop]['unsupported_in'].join(', ')})
-        end
+    property_support = @client_support['css_properties']
+    properties.each do |prop|
+      if property_support.include?(prop) and
+          property_support[prop].include?('support') and
+          property_support[prop]['support'] >= @options[:warn_level]
+        warnings.push({:message => "#{prop} CSS property",
+            :level => WARN_LABEL[property_support[prop]['support']],
+            :clients => property_support[prop]['unsupported_in'].join(', ')})
       end
-
-      return warnings
     end
-  end
 
+    @client_support['attributes'].each do |attribute, data|
+      next unless data['support'] >= @options[:warn_level]
+      if @doc.search("*[@#{attribute}]").length > 0
+        warnings.push({:message => "#{attribute} HTML attribute",
+            :level => WARN_LABEL[data['support']],
+            :clients => data['unsupported_in'].join(', ')})
+      end
+    end
 
+    @client_support['elements'].each do |element, data|
+      next unless data['support'] >= @options[:warn_level]
+      if @doc.search(element).length > 0
+        warnings.push({:message => "#{element} HTML element",
+            :level => WARN_LABEL[data['support']],
+            :clients => data['unsupported_in'].join(', ')})
+      end
+    end
 
+    return warnings
+  end
+end

metadata

@@ -4,9 +4,9 @@ version: !ruby/object:Gem::Version
   prerelease: false
   segments: 
   - 1
-  - 6
-  - 2
-  version: 1.6.2
+  - 7
+  - 0
+  version: 1.7.0
 platform: ruby
 authors: 
 - Alex Dunae
@@ -14,7 +14,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-11-22 00:00:00 -08:00
+date: 2011-12-31 00:00:00 -08:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -43,8 +43,8 @@ dependencies:
         segments: 
         - 1
         - 1
-        - 3
-        version: 1.1.3
+        - 6
+        version: 1.1.6
   type: :runtime
   version_requirements: *id002
 - !ruby/object:Gem::Dependency 
@@ -76,6 +76,9 @@ files:
 - lib/premailer.rb
 - lib/premailer/html_to_plain_text.rb
 - lib/premailer/premailer.rb
+- lib/premailer/adapter.rb
+- lib/premailer/adapter/hpricot.rb
+- lib/premailer/adapter/nokogiri.rb
 - misc/client_support.yaml
 - README.rdoc
 has_rdoc: true