premailer 1.8.6 → 1.8.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 7316d27a2fa70efed484e7c64733a351491f1cea
-  data.tar.gz: 6dd11b25df48ce0e73149ffdc25e6702de94ec21
+  metadata.gz: ceccb9e67d075de22a76d50df38d99481a6811bf
+  data.tar.gz: 475cc5974d20a0b8b69bef467d10de3edff7d821
 SHA512:
-  metadata.gz: 9d5aed590607bd5cdbe7b629424c2b18832e5ee36c5688627216ea660c493edfb92a6f5d2f77e5d78b2b5cb502d2fb7f592adf3e0d624d58ced7b57ac7f6acdb
-  data.tar.gz: 107fb306346e6e35b146b7e327b0e2cc8cca1ded78f75687ed5281e090f08d50e7a9625335fbaffe066496e0c72f121b3389779751933900b7aec356d8b86791
+  metadata.gz: 2591ab12195ae73b9909b945131cdf06382917afde4e4270dc496b937eac87c98cea081c825209a0e4cfd40f8b2202f1953f980d8203cf90afcadec80b139c17
+  data.tar.gz: e34eda9604977da572ef0f712c710547e9e7d965370404a2b1a233a2dd04760a7461c70acf4b40008b72824fe2a38f4cb18bd36d49da794bd61bd41a09096721

data/README.md

@@ -54,7 +54,7 @@ end
 
 ## Ruby Compatibility
 
-Premailer is tested on Ruby 1.8.7, Ruby 1.9.2, Ruby 1.9.3, and Ruby 2.x.0 . It also works on REE. JRuby support is close; contributors are welcome.  Checkout the latest build status on the [Travis CI dashboard](http://travis-ci.org/#!/premailer/premailer).
+Premailer is tested on Ruby 1.8.7, Ruby 1.9.2, Ruby 1.9.3, and Ruby 2.x.0 . It also works on REE. JRuby support is close; contributors are welcome.  Checkout the latest build status on the [Travis CI dashboard](https://travis-ci.org/#!/premailer/premailer).
 
 ## Premailer-specific CSS
 
@@ -66,7 +66,7 @@ Premailer looks for a few CSS attributes that make working with tables a bit eas
 | -premailer-height | Available on `table`, `tr`, `th` and `td` elements |
 | -premailer-cellpadding | Available on `table` elements |
 | -premailer-cellspacing | Available on `table` elements |
-| data-premailer="ignore" | Available on `style` elements. Premailer will ignore these elements entirely. |
+| data-premailer="ignore" | Available on `link` and `style` elements. Premailer will ignore these elements entirely. |
 
 Each of these CSS declarations will be copied to appropriate element's attribute.
 
@@ -93,10 +93,10 @@ A few areas that are particularly in need of love:
 
 ## Credits and code
 
-Thanks to [all the wonderful contributors](https://github.com/alexdunae/premailer/contributors) for their updates.
+Thanks to [all the wonderful contributors](https://github.com/premailer/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.
+and to [Campaign Monitor](https://www.campaignmonitor.com/) for supporting the web interface.
 
 The web interface can be found at [premailer.dialect.ca](http://premailer.dialect.ca).
 

data/bin/premailer

@@ -2,6 +2,5 @@
 
 # This binary used in rubygems environment only as part of installed gem
 
-require 'rubygems'
 require 'premailer/executor'
 

data/lib/premailer/adapter.rb

@@ -4,16 +4,19 @@ class Premailer
   # Manages the adapter classes. Currently supports:
   #
   # * nokogiri
+  # * nokogumbo
   # * hpricot
   module Adapter
 
     autoload :Hpricot, 'premailer/adapter/hpricot'
     autoload :Nokogiri, 'premailer/adapter/nokogiri'
+    autoload :Nokogumbo, 'premailer/adapter/nokogumbo'
 
     # adapter to required file mapping.
     REQUIREMENT_MAP = [
-      ["hpricot",  :hpricot],
       ["nokogiri", :nokogiri],
+      ["nokogumbo", :nokogumbo],
+      ["hpricot",  :hpricot],
     ]
 
     # Returns the adapter to use.
@@ -28,8 +31,9 @@ class Premailer
     # then checks to see which are installed if none are loaded.
     # @raise [RuntimeError] unless suitable adapter found.
     def self.default
-      return :hpricot  if defined?(::Hpricot)
       return :nokogiri if defined?(::Nokogiri)
+      return :nokogumbo if defined?(::Nokogumbo)
+      return :hpricot  if defined?(::Hpricot)
 
       REQUIREMENT_MAP.each do |(library, adapter)|
         begin

data/lib/premailer/adapter/hpricot.rb

@@ -4,6 +4,13 @@ class Premailer
   module Adapter
     # Hpricot adapter
     module Hpricot
+      def self.included(base)
+        warn <<eos
+[DEPRECATED] Premailer's Hpricot adapter will be removed with the next major \
+release. Change your :adapter option to :nokogiri or remove the `hpricot` gem \
+from your application to silence this warning. (Called from #{caller[2]})
+eos
+      end
 
       # Merge CSS into the HTML document.
       # @return [String] HTML.
@@ -75,12 +82,15 @@ class Premailer
           # Perform style folding
           merged = CssParser.merge(declarations)
           merged.expand_shorthand!
-          merged.create_shorthand!
+          merged.create_shorthand! if @options[:create_shorthands]
 
           # Duplicate CSS attributes as HTML attributes
-          if Premailer::RELATED_ATTRIBUTES.has_key?(el.name)
+          if Premailer::RELATED_ATTRIBUTES.has_key?(el.name) && @options[:css_to_attributes]
             Premailer::RELATED_ATTRIBUTES[el.name].each do |css_att, html_att|
               el[html_att] = merged[css_att].gsub(/url\('(.*)'\)/,'\1').gsub(/;$|\s*!important/, '').strip if el[html_att].nil? and not merged[css_att].empty?
+              merged.instance_variable_get("@declarations").tap do |declarations|
+                declarations.delete(css_att)
+              end
             end
           end
 

data/lib/premailer/adapter/nokogiri.rb

@@ -17,7 +17,6 @@ class Premailer
         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, media_types|
           # Save un-mergable rules separately
@@ -74,19 +73,22 @@ class Premailer
           merged.expand_shorthand!
 
           # Duplicate CSS attributes as HTML attributes
-          if Premailer::RELATED_ATTRIBUTES.has_key?(el.name)
+          if Premailer::RELATED_ATTRIBUTES.has_key?(el.name) && @options[:css_to_attributes]
             Premailer::RELATED_ATTRIBUTES[el.name].each do |css_att, html_att|
               el[html_att] = merged[css_att].gsub(/url\(['|"](.*)['|"]\)/, '\1').gsub(/;$|\s*!important/, '').strip if el[html_att].nil? and not merged[css_att].empty?
+              merged.instance_variable_get("@declarations").tap do |declarations|
+                declarations.delete(css_att)
+              end
             end
           end
-
           # Collapse multiple rules into one as much as possible.
-          merged.create_shorthand!
+          merged.create_shorthand! if @options[:create_shorthands]
 
           # write the inline STYLE attribute
-          attributes = Premailer.escape_string(merged.declarations_to_s).split(';').map(&:strip)
+          # split by ';' but ignore those in brackets
+          attributes = Premailer.escape_string(merged.declarations_to_s).split(/;(?![^(]*\))/).map(&:strip)
           attributes = attributes.map { |attr| [attr.split(':').first, attr] }.sort_by { |pair| pair.first }.map { |pair| pair[1] }
-          el['style'] = attributes.join('; ')
+          el['style'] = attributes.join('; ') + ";"
         end
 
         doc = write_unmergable_css_rules(doc, @unmergable_rules)

data/lib/premailer/adapter/nokogumbo.rb

@@ -0,0 +1,243 @@
+require 'nokogumbo'
+
+class Premailer
+  module Adapter
+    # Nokogiri adapter
+    module Nokogumbo
+
+      # Merge CSS into the HTML document.
+      #
+      # @return [String] an HTML.
+      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, media_types|
+          # 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 Premailer.is_media_query?(media_types) || selector =~ Premailer::RE_UNMERGABLE_SELECTORS
+            @unmergable_rules.add_rule_set!(CssParser::RuleSet.new(selector, declaration), media_types) unless @options[:preserve_styles]
+          else
+            begin
+              if selector =~ Premailer::RE_RESET_SELECTORS
+                # this is in place to preserve the MailChimp CSS reset: http://github.com/mailchimp/Email-Blueprints/
+                # however, this doesn't mean for testing pur
+                @unmergable_rules.add_rule_set!(CssParser::RuleSet.new(selector, declaration)) unless !@options[:preserve_reset]
+              end
+
+              # 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
+
+        # Remove script tags
+        if @options[:remove_scripts]
+          doc.search("script").remove
+        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) && @options[:css_to_attributes]
+            Premailer::RELATED_ATTRIBUTES[el.name].each do |css_att, html_att|
+              el[html_att] = merged[css_att].gsub(/url\(['|"](.*)['|"]\)/, '\1').gsub(/;$|\s*!important/, '').strip if el[html_att].nil? and not merged[css_att].empty?
+              merged.instance_variable_get("@declarations").tap do |declarations|
+                declarations.delete(css_att)
+              end
+            end
+          end
+          # Collapse multiple rules into one as much as possible.
+          merged.create_shorthand! if @options[:create_shorthands]
+
+          # write the inline STYLE attribute
+          attributes = Premailer.escape_string(merged.declarations_to_s).split(';').map(&:strip)
+          attributes = attributes.map { |attr| [attr.split(':').first, attr] }.sort_by { |pair| pair.first }.map { |pair| pair[1] }
+          el['style'] = attributes.join('; ') + ";"
+        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
+
+        if @options[:reset_contenteditable]
+          doc.search('*[@contenteditable]').each do |el|
+            el.remove_attribute('contenteditable')
+          end
+        end
+
+        @processed_doc = doc
+        if is_xhtml?
+          # we don't want to encode carriage returns
+          @processed_doc.to_xhtml(:encoding => @options[:output_encoding]).gsub(/&\#(xD|13);/i, "\r")
+        else
+          @processed_doc.to_html(:encoding => @options[:output_encoding])
+        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.
+      #
+      # @return [::Nokogiri::XML] a document.
+      def write_unmergable_css_rules(doc, unmergable_rules) # :nodoc:
+        styles = unmergable_rules.to_s
+
+        unless styles.empty?
+          style_tag = "<style type=\"text/css\">\n#{styles}</style>"
+          unless (body = doc.search('body')).empty?
+            if doc.at_css('body').children && !doc.at_css('body').children.empty?
+              doc.at_css('body').children.before(::Nokogiri::XML.fragment(style_tag))
+            else
+              doc.at_css('body').add_child(::Nokogiri::XML.fragment(style_tag))
+            end
+          else
+            doc.inner_html = style_tag += doc.inner_html
+          end
+        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.
+      #
+      # @return [String] a plain text.
+      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
+
+      # Gets the original HTML as a string.
+      # @return [String] HTML.
+      def to_s
+        if is_xhtml?
+          @doc.to_xhtml(:encoding => nil)
+        else
+          @doc.to_html(:encoding => nil)
+        end
+      end
+
+      # Load the HTML file and convert it into an Nokogiri document.
+      #
+      # @return [::Nokogiri::XML] a 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
+
+        # Handle HTML entities
+        if @options[:replace_html_entities] == true and thing.is_a?(String)
+          HTML_ENTITIES.map do |entity, replacement|
+            thing.gsub! entity, replacement
+          end
+        end
+        # Default encoding is ASCII-8BIT (binary) per http://groups.google.com/group/nokogiri-talk/msg/0b81ef0dc180dc74
+        # However, we really don't want to hardcode this. ASCII-8BIT should be the default, but not the only option.
+        if thing.is_a?(String) and RUBY_VERSION =~ /1.9/
+          thing = thing.force_encoding(@options[:input_encoding]).encode!
+          doc = ::Nokogiri::HTML5(thing)
+        else
+          default_encoding = RUBY_PLATFORM == 'java' ? nil : 'BINARY'
+          doc = ::Nokogiri::HTML5(thing)
+        end
+
+        # Fix for removing any CDATA tags from both style and script tags inserted per
+        # https://github.com/sparklemotion/nokogiri/issues/311 and
+        # https://github.com/premailer/premailer/issues/199
+        %w(style script).each do |tag|
+          doc.search(tag).children.each do |child|
+            child.swap(child.text()) if child.cdata?
+          end
+        end
+
+        doc
+      end
+
+    end
+  end
+end

data/lib/premailer/html_to_plain_text.rb

@@ -31,17 +31,11 @@ module HtmlToPlainText
     txt.gsub!(/<img.+?alt=\'([^\']*)\'[^>]*\>/i, '\1')
 
     # links
-    txt.gsub!(/<a\s.*?href=\"(mailto:)?([^\"]*)\"[^>]*>((.|\s)*?)<\/a>/i) do |s|
-      if $3.empty?
-        ''
-      else
-        $3.strip + ' ( ' + $2.strip + ' )'
-      end
-    end
-
-    txt.gsub!(/<a\s.*?href='(mailto:)?([^\']*)\'[^>]*>((.|\s)*?)<\/a>/i) do |s|
+    txt.gsub!(/<a\s.*?href=["'](mailto:)?([^"']*)["'][^>]*>((.|\s)*?)<\/a>/i) do |s|
       if $3.empty?
         ''
+      elsif $3.strip.downcase == $2.strip.downcase
+        $3.strip
       else
         $3.strip + ' ( ' + $2.strip + ' )'
       end
@@ -92,25 +86,25 @@ module HtmlToPlainText
     he = HTMLEntities.new
     txt = he.decode(txt)
 
+    # no more than two consecutive spaces
+    txt.gsub!(/ {2,}/, " ")
+
     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!(/[ \t]*\302\240+[ \t]*/, " ") # non-breaking spaces -> spaces
     txt.gsub!(/\n[ \t]+/, "\n") # space at start of lines
     txt.gsub!(/[ \t]+\n/, "\n") # space at end of lines
 
     # no more than two consecutive newlines
     txt.gsub!(/[\n]{3,}/, "\n\n")
 
-    # no more than two consecutive spaces
-    txt.gsub!(/ {2,}/, " ")
-
     # the word messes up the parens
-    txt.gsub!(/\([ \n](http[^)]+)[\n ]\)/) do |s|
-      "( " + $1 + " )"
+    txt.gsub!(/\(([ \n])(http[^)]+)([\n ])\)/) do |s|
+      ($1 == "\n" ? $1 : '' ) + '( ' + $2 + ' )' + ($3 == "\n" ? $1 : '' )
     end
 
     txt.strip

data/lib/premailer/premailer.rb

@@ -166,6 +166,7 @@ class Premailer
   # @option options [Boolean] :preserve_reset Whether to preserve styles associated with the MailChimp reset code. Default is true.
   # @option options [Boolean] :with_html_string Whether the html param should be treated as a raw string. Default is false.
   # @option options [Boolean] :verbose Whether to print errors and warnings to <tt>$stderr</tt>.  Default is false.
+  # @option options [Boolean] :io_exceptions Throws exceptions on I/O errors.
   # @option options [Boolean] :include_link_tags Whether to include css from <tt>link rel=stylesheet</tt> tags.  Default is true.
   # @option options [Boolean] :include_style_tags Whether to include css from <tt>style</tt> tags.  Default is true.
   # @option options [String] :input_encoding Manually specify the source documents encoding. This is a good idea. Default is ASCII-8BIT.
@@ -173,6 +174,7 @@ class Premailer
   # @option options [Boolean] :escape_url_attributes URL Escapes href, src, and background attributes on elements. Default is true.
   # @option options [Symbol] :adapter Which HTML parser to use, either <tt>:nokogiri</tt> or <tt>:hpricot</tt>.  Default is <tt>:hpricot</tt>.
   # @option options [String] :output_encoding Output encoding option for Nokogiri adapter. Should be set to "US-ASCII" to output HTML entities instead of Unicode characters.
+  # @option options [Boolean] :create_shorthands Combine several properties into a shorthand one, e.g. font: style weight size. Default is true.
   def initialize(html, options = {})
     @options = {:warn_level => Warnings::SAFE,
                 :line_length => 65,
@@ -199,6 +201,7 @@ class Premailer
                 :replace_html_entities => false,
                 :escape_url_attributes => true,
                 :unescaped_ampersand => false,
+                :create_shorthands => true,
                 :adapter => Adapter.use,
                 }.merge(options)
 
@@ -260,7 +263,9 @@ protected
       end
 
       load_css_from_string(css_block)
-    rescue; end
+    rescue => e
+      raise e if @options[:io_exceptions]
+    end
   end
 
   def load_css_from_string(css_string)
@@ -282,11 +287,7 @@ protected
 
     # Load CSS included in <tt>style</tt> and <tt>link</tt> tags from an HTML document.
   def load_css_from_html! # :nodoc:
-    if (@options[:adapter] == :nokogiri)
-      tags = @doc.search("link[@rel='stylesheet']:not([@data-premailer='ignore'])", "//style[not(contains(@data-premailer,'ignore'))]")
-    else
-      tags = @doc.search("link[@rel='stylesheet']:not([@data-premailer='ignore']), style:not([@data-premailer='ignore'])")
-    end
+    tags = @doc.search("link[@rel='stylesheet']:not([@data-premailer='ignore']), style:not([@data-premailer='ignore'])")
     if tags
       tags.each do |tag|
         if tag.to_s.strip =~ /^\<link/i && tag.attributes['href'] && media_type_ok?(tag.attributes['media']) && @options[:include_link_tags]
@@ -298,10 +299,10 @@ protected
               link_uri = tag.attributes['href'].to_s.sub(@base_url.to_s, '')
             else
               link_uri = File.join(File.dirname(@html_file), tag.attributes['href'].to_s.sub!(@base_url.to_s, ''))
-              # if the file does not exist locally, try to grab the remote reference
-              unless File.exists?(link_uri)
-                link_uri = Premailer.resolve_link(tag.attributes['href'].to_s, @html_file)
-              end
+            end
+            # if the file does not exist locally, try to grab the remote reference
+            unless File.exists?(link_uri)
+              link_uri = Premailer.resolve_link(tag.attributes['href'].to_s, @html_file)
             end
           else
             link_uri = tag.attributes['href'].to_s