html2haml 1.0.1 → 2.0.0.beta.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: d8e73acdd5aa2e60c2475d9fd27bfd5745b77518
+  data.tar.gz: 368387da16cf45f3bd9149e5cd9f8a5b9fe9e1ec
+SHA512:
+  metadata.gz: 92c049a3f1c46df05f913af77bd2d6e9a2439e2c3d32ab28db9d351a4c22b95da168d5ad1ce29bf7b1333bd0c3d5e044266d78dbf9a950f7b3a45637e33b2724
+  data.tar.gz: c5314e4c81f28df2be455ab26b4c58ca1de9ed88c8e7ff4e2eeee0a3b8210160db86df8c5ed6ff6adad4257b54366a4bcc6fb3e72527f900af2ff8106de6cf08

data/.travis.yml

@@ -1,8 +1,7 @@
 rvm:
-  - 1.8.7
   - 1.9.3
-  - jruby-18mode
-  - rbx-18mode
+  - 2.0.0
+  - rbx-19mode
 
 gemfile:
   - Gemfile

data/.yardopts

@@ -0,0 +1 @@
+--markup markdown

data/Changelog.markdown

@@ -1,5 +1,25 @@
 # HTML2Haml Changelog
 
+## 2.0.0
+
+* Switch to Nokogiri for XML parsing.
+  (thanks to [Stefan Natchev](https://github.com/snatchev) and [Norman
+  Clarke](https://github.com/norman))
+
+* Add Ruby 2.0 support.
+  (thanks to [Yasuharu Ozaki](https://github.com/yasuoza))
+
+* Add option to use Ruby 1.9-style attributes when possible.
+  (thanks to [Yoshinori Kawasaki](https://github.com/luvtechno) and
+  [Alexander Egorov](https://github.com/qatsi))
+
+* Updated dependency versions.
+
+* Removed some deprecated configuration flags.
+
+* Move the internal HTML class from the Haml namespace into the Html2haml
+  namespace.
+
 ## 1.0.1
 
 Rescue from `RubyParser::SyntaxError` in check for valid ruby.

data/README.md

@@ -1,6 +1,9 @@
 # Html2haml
 
-Converts HTML to Haml.
+[![Build Status](https://travis-ci.org/haml/html2haml.png?branch=master)](https://travis-ci.org/haml/html2haml)
+
+Html2haml, not surprisingly, converts HTML to Haml. It works on HTML with
+embedded ERB tags as well as plain old HTML.
 
 ## Installation
 
@@ -18,6 +21,33 @@ Or install it yourself as:
 
 ## Usage
 
+
+### To convert a project from .erb to .haml
+
+If your system has `sed` and `xargs` available and none of your .erb file names
+have whitespace in them, you can convert all your templates like so:
+
+    find . -name \*.erb -print | sed 'p;s/.erb$/.haml/' | xargs -n2 html2haml
+
+If some of your file names have whitespace or you need finer-grained control
+over the process, you can convert your files using `gsed` or multi-line script
+techniques discussed [here](http://stackoverflow.com/questions/17576814/).
+
+
+### Documentation
+
+#### About version 2.0
+
+Html2haml 2.0 differs from 1.x primarily in that it uses Nokgiri as its HTML
+parser rather than Hpricot. At the current time however, there are some
+problems running Html2haml 2.0 on JRuby due to differences in the way the Java
+version of Nokogiri parses HTML. If you are using JRuby you may wish to run
+HTML2Haml on MRI or use a 1.x version until these problems have been resolved.
+
+#### Options
+
+Here are the options currently available to Html2haml:
+
 See `html2haml --help`:
 
     Usage: html2haml [options] [INPUT] [OUTPUT]
@@ -27,10 +57,8 @@ See `html2haml --help`:
     Options:
         -e, --erb                        Parse ERb tags.
             --no-erb                     Don't parse ERb tags.
-        -r, --rhtml                      Deprecated; same as --erb.
-            --no-rhtml                   Deprecated; same as --no-erb.
-        -x, --xhtml                      Parse the input using the more strict XHTML parser.
             --html-attributes            Use HTML style attributes instead of Ruby hash style.
+            --ruby19-attributes          Use Ruby 1.9-style attributes when possible.
         -E ex[:in]                       Specify the default external and internal character encodings.
         -s, --stdin                      Read input from standard input instead of an input file
             --trace                      Show a full traceback on error

data/Rakefile

@@ -16,7 +16,7 @@ task :set_coverage_env do
   ENV["COVERAGE"] = "true"
 end
 
-desc "Run Simplecov (only works on 1.9)"
+desc "Run Simplecov"
 task :coverage => [:set_coverage_env, :test]
 
 gemspec = File.expand_path("../html2haml.gemspec", __FILE__)

data/html2haml.gemspec

@@ -2,8 +2,8 @@
 require File.expand_path('../lib/html2haml/version', __FILE__)
 
 Gem::Specification.new do |gem|
-  gem.authors       = ["Norman Clarke"]
-  gem.email         = ["norman@njclarke.com"]
+  gem.authors       = ["Norman Clarke", "Stefan Natchev"]
+  gem.email         = ["norman@njclarke.com", "stefan.natchev@gmail.com"]
   gem.description   = %q{Converts HTML into Haml}
   gem.summary       = %q{Converts HTML into Haml}
   gem.homepage      = "http://haml.info"
@@ -15,10 +15,12 @@ Gem::Specification.new do |gem|
   gem.require_paths = ["lib"]
   gem.version       = Html2haml::VERSION
 
-  gem.add_dependency 'hpricot', '~> 0.8.6'
+  gem.required_ruby_version = '>= 1.9.2'
+
+  gem.add_dependency 'nokogiri', '~> 1.6.0'
   gem.add_dependency 'erubis', '~> 2.7.0'
-  gem.add_dependency 'ruby_parser', '~> 3.1.1'
-  gem.add_dependency 'haml', '>= 4.0.0.rc.1'
+  gem.add_dependency 'ruby_parser', '~> 3.2.1'
+  gem.add_dependency 'haml', '~> 4.0.0'
   gem.add_development_dependency 'simplecov', '~> 0.7.1'
   gem.add_development_dependency 'minitest', '~> 4.4.0'
   gem.add_development_dependency 'rake'

data/lib/html2haml/exec.rb

@@ -3,9 +3,12 @@ require 'fileutils'
 require 'rbconfig'
 
 module Html2haml
-  # This module handles the various Haml executables (`haml` and `haml-convert`).
+  # This module handles the Html2haml executable.
   module Exec
-    # An abstract class that encapsulates the executable code for all three executables.
+    # An abstract class that encapsulates the executable code for the Html2haml executable.
+    # It's split into a base class and a subclass for historic reasons: this previously
+    # was used by all the executables in the Haml project, before Html2haml was moved
+    # into its own gem.
     class Generic
       # @param args [Array<String>] The command-line arguments
       def initialize(args)
@@ -208,28 +211,18 @@ END
           @options[:no_erb] = true
         end
 
-        opts.on('-r', '--rhtml', 'Deprecated; same as --erb.') do
-          @module_opts[:erb] = true
-        end
-
-        opts.on('--no-rhtml', "Deprecated; same as --no-erb.") do
-          @options[:no_erb] = true
-        end
-
-        opts.on('-x', '--xhtml', 'Parse the input using the more strict XHTML parser.') do
-          @module_opts[:xhtml] = true
-        end
-
         opts.on("--html-attributes", "Use HTML style attributes instead of Ruby hash style.") do
           @module_opts[:html_style_attributes] = true
         end
 
-        unless RUBY_VERSION < "1.9"
-          opts.on('-E ex[:in]', 'Specify the default external and internal character encodings.') do |encoding|
-            external, internal = encoding.split(':')
-            Encoding.default_external = external if external && !external.empty?
-            Encoding.default_internal = internal if internal && !internal.empty?
-          end
+        opts.on("--ruby19-attributes", "Use Ruby 1.9-style attributes when possible.") do
+          @module_opts[:ruby19_style_attributes] = true
+        end
+
+        opts.on('-E ex[:in]', 'Specify the default external and internal character encodings.') do |encoding|
+          external, internal = encoding.split(':')
+          Encoding.default_external = external if external && !external.empty?
+          Encoding.default_internal = internal if internal && !internal.empty?
         end
 
         super
@@ -248,7 +241,7 @@ END
         @module_opts[:erb] ||= input.respond_to?(:path) && input.path =~ /\.(rhtml|erb)$/
         @module_opts[:erb] &&= @options[:no_erb] != false
 
-        output.write(::Haml::HTML.new(input, @module_opts).render)
+        output.write(HTML.new(input, @module_opts).render)
       rescue ::Haml::Error => e
         raise "#{e.is_a?(::Haml::SyntaxError) ? "Syntax error" : "Error"} on line " +
           "#{get_line e}: #{e.message}"

data/lib/html2haml/html.rb

@@ -1,116 +1,119 @@
 require 'cgi'
-require 'hpricot'
+require 'nokogiri'
 require 'html2haml/html/erb'
 
-# Haml monkeypatches various Hpricot classes
+# Haml monkeypatches various Nokogiri classes
 # to add methods for conversion to Haml.
 # @private
-module Hpricot
-  # @see Hpricot
-  module Node
-    # Whether this node has already been converted to Haml.
-    # Only used for text nodes and elements.
-    #
-    # @return [Boolean]
-    attr_accessor :converted_to_haml
-
-    # Returns the Haml representation of the given node.
-    #
-    # @param tabs [Fixnum] The indentation level of the resulting Haml.
-    # @option options (see Haml::HTML#initialize)
-    def to_haml(tabs, options)
-      return "" if converted_to_haml || to_s.strip.empty?
-      text = uninterp(self.to_s)
-      node = next_node
-      while node.is_a?(::Hpricot::Elem) && node.name == "haml:loud"
-        node.converted_to_haml = true
-        text << '#{' <<
-          CGI.unescapeHTML(node.inner_text).gsub(/\n\s*/, ' ').strip << '}'
-
-        if node.next_node.is_a?(::Hpricot::Text)
-          node = node.next_node
-          text << uninterp(node.to_s)
+module Nokogiri
+
+  module XML
+    # @see Nokogiri
+    class Node
+      # Whether this node has already been converted to Haml.
+      # Only used for text nodes and elements.
+      #
+      # @return [Boolean]
+      attr_accessor :converted_to_haml
+
+      # Returns the Haml representation of the given node.
+      #
+      # @param tabs [Fixnum] The indentation level of the resulting Haml.
+      # @option options (see Html2haml::HTML#initialize)
+      def to_haml(tabs, options)
+        return "" if converted_to_haml || to_s.strip.empty?
+        text = uninterp(self.to_s)
+        node = next_sibling
+        while node.is_a?(::Nokogiri::XML::Element) && node.name == "haml_loud"
           node.converted_to_haml = true
-        end
+          text << '#{' <<
+            CGI.unescapeHTML(node.inner_text).gsub(/\n\s*/, ' ').strip << '}'
+
+          if node.next_sibling.is_a?(::Nokogiri::XML::Text)
+            node = node.next_sibling
+            text << uninterp(node.to_s)
+            node.converted_to_haml = true
+          end
 
-        node = node.next_node
+          node = node.next_sibling
+        end
+        return parse_text_with_interpolation(text, tabs)
       end
-      return parse_text_with_interpolation(text, tabs)
-    end
 
-    private
-
-    def erb_to_interpolation(text, options)
-      return text unless options[:erb]
-      text = CGI.escapeHTML(uninterp(text))
-      %w[<haml:loud> </haml:loud>].each {|str| text.gsub!(CGI.escapeHTML(str), str)}
-      ::Hpricot::XML(text).children.inject("") do |str, elem|
-        if elem.is_a?(::Hpricot::Text)
-          str + CGI.unescapeHTML(elem.to_s)
-        else # <haml:loud> element
-          str + '#{' + CGI.unescapeHTML(elem.innerText.strip) + '}'
+      private
+
+      def erb_to_interpolation(text, options)
+        return text unless options[:erb]
+        text = CGI.escapeHTML(uninterp(text))
+        %w[<haml_loud> </haml_loud>].each {|str| text.gsub!(CGI.escapeHTML(str), str)}
+        ::Nokogiri::XML.fragment(text).children.inject("") do |str, elem|
+          if elem.is_a?(::Nokogiri::XML::Text)
+            str + CGI.unescapeHTML(elem.to_s)
+          else # <haml_loud> element
+            str + '#{' + CGI.unescapeHTML(elem.inner_text.strip) + '}'
+          end
         end
       end
-    end
 
-    def tabulate(tabs)
-      '  ' * tabs
-    end
+      def tabulate(tabs)
+        '  ' * tabs
+      end
 
-    def uninterp(text)
-      text.gsub('#{', '\#{') #'
-    end
+      def uninterp(text)
+        text.gsub('#{', '\#{') #'
+      end
 
-    def attr_hash
-      attributes.to_hash
-    end
+      def attr_hash
+        Hash[attributes.map {|k, v| [k.to_s, v.to_s]}]
+      end
 
-    def parse_text(text, tabs)
-      parse_text_with_interpolation(uninterp(text), tabs)
-    end
+      def parse_text(text, tabs)
+        parse_text_with_interpolation(uninterp(text), tabs)
+      end
 
-    def parse_text_with_interpolation(text, tabs)
-      text.strip!
-      return "" if text.empty?
+      def parse_text_with_interpolation(text, tabs)
+        text.strip!
+        return "" if text.empty?
 
-      text.split("\n").map do |line|
-        line.strip!
-        "#{tabulate(tabs)}#{'\\' if Haml::Parser::SPECIAL_CHARACTERS.include?(line[0])}#{line}\n"
-      end.join
+        text.split("\n").map do |line|
+          line.strip!
+          "#{tabulate(tabs)}#{'\\' if Haml::Parser::SPECIAL_CHARACTERS.include?(line[0])}#{line}\n"
+        end.join
+      end
     end
   end
 end
 
 # @private
-HAML_TAGS = %w[haml:block haml:loud haml:silent]
-
-HAML_TAGS.each do |t|
-  Hpricot::ElementContent[t] = {}
-  Hpricot::ElementContent.keys.each do |key|
-    Hpricot::ElementContent[t][key.hash] = true
-  end
-end
-
-Hpricot::ElementContent.keys.each do |k|
-  HAML_TAGS.each do |el|
-    val = Hpricot::ElementContent[k]
-    val[el.hash] = true if val.is_a?(Hash)
-  end
-end
-
-module Haml
+HAML_TAGS = %w[haml_block haml_loud haml_silent]
+#
+# HAML_TAGS.each do |t|
+#   Nokogiri::XML::ElementContent[t] = {}
+#   Nokogiri::XML::ElementContent.keys.each do |key|
+#     Nokogiri::XML::ElementContent[t][key.hash] = true
+#   end
+# end
+#
+# Nokogiri::XML::ElementContent.keys.each do |k|
+#   HAML_TAGS.each do |el|
+#     val = Nokogiri::XML::ElementContent[k]
+#     val[el.hash] = true if val.is_a?(Hash)
+#   end
+# end
+
+module Html2haml
   # Converts HTML documents into Haml templates.
-  # Depends on [Hpricot](http://github.com/whymirror/hpricot) for HTML parsing.
-  # If ERB conversion is being used, also depends on 
+  # Depends on [Nokogiri](http://nokogiri.org/) for HTML parsing.
+  # If ERB conversion is being used, also depends on
   # [Erubis](http://www.kuwata-lab.com/erubis) to parse the ERB
   # and [ruby_parser](http://parsetree.rubyforge.org/) to parse the Ruby code.
   #
   # Example usage:
   #
-  #     Haml::HTML.new("<a href='http://google.com'>Blat</a>").render
+  #     HTML.new("<a href='http://google.com'>Blat</a>").render
   #       #=> "%a{:href => 'http://google.com'} Blat"
   class HTML
-    # @param template [String, Hpricot::Node] The HTML template to convert
+    # @param template [String, Nokogiri::Node] The HTML template to convert
     # @option options :erb [Boolean] (false) Whether or not to parse
     #   ERB's `<%= %>` and `<% %>` into Haml's `=` and `-`
     # @option options :xhtml [Boolean] (false) Whether or not to parse
@@ -118,7 +121,7 @@ module Haml
     def initialize(template, options = {})
       @options = options
 
-      if template.is_a? Hpricot::Node
+      if template.is_a? Nokogiri::XML::Node
         @template = template
       else
         if template.is_a? IO
@@ -132,8 +135,23 @@ module Haml
           template = ERB.compile(template)
         end
 
-        method = @options[:xhtml] ? Hpricot.method(:XML) : method(:Hpricot)
-        @template = method.call(template.gsub('&', '&amp;'))
+        if template =~ /^\s*<!DOCTYPE|<html/i
+          return @template = Nokogiri.HTML(template)
+        end
+
+        @template = Nokogiri::HTML.fragment(template)
+
+        #detect missplaced head or body tag
+        #XML_HTML_STRUCURE_ERROR : 800
+        if @template.errors.any? { |e| e.code == 800 }
+          return @template = Nokogiri.HTML(template).at('/html').children
+        end
+
+        #in order to support CDATA in HTML (which is invalid) try using the XML parser
+        # we can detect this when libxml returns error code XML_ERR_NAME_REQUIRED : 68
+        if @template.errors.any? { |e| e.code == 68 }
+          return @template = Nokogiri::XML.fragment(template)
+        end
       end
     end
 
@@ -146,42 +164,64 @@ module Haml
 
     TEXT_REGEXP = /^(\s*).*$/
 
-    # @see Hpricot
+
+    # @see Nokogiri
     # @private
-    class ::Hpricot::Doc
-      # @see Haml::HTML::Node#to_haml
+    class ::Nokogiri::XML::Document
+      # @see Html2haml::HTML::Node#to_haml
       def to_haml(tabs, options)
         (children || []).inject('') {|s, c| s << c.to_haml(0, options)}
       end
     end
 
-    # @see Hpricot
+    class ::Nokogiri::XML::DocumentFragment
+      # @see Html2haml::HTML::Node#to_haml
+      def to_haml(tabs, options)
+        (children || []).inject('') {|s, c| s << c.to_haml(0, options)}
+      end
+    end
+
+    class ::Nokogiri::XML::NodeSet
+      # @see Html2haml::HTML::Node#to_haml
+      def to_haml(tabs, options)
+        self.inject('') {|s, c| s << c.to_haml(tabs, options)}
+      end
+    end
+
+    # @see Nokogiri
     # @private
-    class ::Hpricot::XMLDecl
-      # @see Haml::HTML::Node#to_haml
+    class ::Nokogiri::XML::ProcessingInstruction
+      # @see Html2haml::HTML::Node#to_haml
       def to_haml(tabs, options)
         "#{tabulate(tabs)}!!! XML\n"
       end
     end
 
-    # @see Hpricot
+    # @see Nokogiri
     # @private
-    class ::Hpricot::CData
-      # @see Haml::HTML::Node#to_haml
+    class ::Nokogiri::XML::CDATA
+      # @see Html2haml::HTML::Node#to_haml
       def to_haml(tabs, options)
         content = parse_text_with_interpolation(
           erb_to_interpolation(self.content, options), tabs + 1)
         "#{tabulate(tabs)}:cdata\n#{content}"
       end
+
+      # removes the start and stop markers for cdata
+      def content_without_cdata_tokens
+        content.
+          gsub(/^\s*<!\[CDATA\[\n/,"").
+          gsub(/^\s*\]\]>\n/, "")
+      end
     end
 
-    # @see Hpricot
+    # @see Nokogiri
     # @private
-    class ::Hpricot::DocType
-      # @see Haml::HTML::Node#to_haml
+    class ::Nokogiri::XML::DTD
+      # @see Html2haml::HTML::Node#to_haml
       def to_haml(tabs, options)
-        attrs = public_id.nil? ? ["", "", ""] :
-          public_id.scan(/DTD\s+([^\s]+)\s*([^\s]*)\s*([^\s]*)\s*\/\//)[0]
+        attrs = external_id.nil? ? ["", "", ""] :
+          external_id.scan(/DTD\s+([^\s]+)\s*([^\s]*)\s*([^\s]*)\s*\/\//)[0]
         raise Haml::SyntaxError.new("Invalid doctype") if attrs == nil
 
         type, version, strictness = attrs.map { |a| a.downcase }
@@ -205,10 +245,10 @@ module Haml
       end
     end
 
-    # @see Hpricot
+    # @see Nokogiri
     # @private
-    class ::Hpricot::Comment
-      # @see Haml::HTML::Node#to_haml
+    class ::Nokogiri::XML::Comment
+      # @see Html2haml::HTML::Node#to_haml
       def to_haml(tabs, options)
         content = self.content
         if content =~ /\A(\[[^\]]+\])>(.*)<!\[endif\]\z/m
@@ -224,26 +264,26 @@ module Haml
       end
     end
 
-    # @see Hpricot
+    # @see Nokogiri
     # @private
-    class ::Hpricot::Elem
-      # @see Haml::HTML::Node#to_haml
+    class ::Nokogiri::XML::Element
+      # @see Html2haml::HTML::Node#to_haml
       def to_haml(tabs, options)
         return "" if converted_to_haml
         if name == "script" &&
-            (attr_hash['type'].nil? || attr_hash['type'] == "text/javascript") &&
+            (attr_hash['type'].nil? || attr_hash['type'].to_s == "text/javascript") &&
             (attr_hash.keys - ['type']).empty?
           return to_haml_filter(:javascript, tabs, options)
         elsif name == "style" &&
-            (attr_hash['type'].nil? || attr_hash['type'] == "text/css") &&
+            (attr_hash['type'].nil? || attr_hash['type'].to_s == "text/css") &&
             (attr_hash.keys - ['type']).empty?
           return to_haml_filter(:css, tabs, options)
         end
 
         output = tabulate(tabs)
-        if options[:erb] && name[0...5] == 'haml:'
-          case name[5..-1]
-          when "loud"
+        if options[:erb] && HAML_TAGS.include?(name)
+          case name
+          when "haml_loud"
             lines = CGI.unescapeHTML(inner_text).split("\n").
               map {|s| s.rstrip}.reject {|s| s.strip.empty?}
             lines.first.gsub!(/^[ \t]*/, "= ")
@@ -261,18 +301,18 @@ module Haml
               length = lines.map {|s| s.size}.max + 1
               lines.map! {|s| "%#{-length}s|" % s}
 
-              if next_sibling && next_sibling.is_a?(Hpricot::Elem) && next_sibling.name == "haml:loud" &&
+              if next_sibling && next_sibling.is_a?(Nokogiri::XML::Element) && next_sibling.name == "haml_loud" &&
                   next_sibling.inner_text.split("\n").reject {|s| s.strip.empty?}.size > 1
                 lines << "-#"
               end
             end
             return lines.map {|s| output + s + "\n"}.join
-          when "silent"
+          when "haml_silent"
             return CGI.unescapeHTML(inner_text).split("\n").map do |line|
               next "" if line.strip.empty?
               "#{output}- #{line.strip}\n"
             end.join
-          when "block"
+          when "haml_block"
             return render_children("", tabs, options)
           end
         end
@@ -286,21 +326,24 @@ module Haml
             output << "= succeed #{self.next.content.slice!(/\A[^\s]+/).dump} do\n"
             tabs += 1
             output << tabulate(tabs)
+            #empty the text node since it was inserted into the block
+            self.next.content = ""
           end
         end
 
-        output << "%#{name}" unless name == 'div' &&
+        output << "%#{name}" unless name.to_s == 'div' &&
           (static_id?(options) ||
            static_classname?(options) &&
-           attr_hash['class'].split(' ').any?(&method(:haml_css_attr?)))
+           attr_hash['class'].to_s.split(' ').any?(&method(:haml_css_attr?)))
 
         if attr_hash
+
           if static_id?(options)
-            output << "##{attr_hash['id']}"
+            output << "##{attr_hash['id'].to_s}"
             remove_attribute('id')
           end
           if static_classname?(options)
-            leftover = attr_hash['class'].split(' ').reject do |c|
+            leftover = attr_hash['class'].to_s.split(' ').reject do |c|
               next unless haml_css_attr?(c)
               output << ".#{c}"
             end
@@ -311,21 +354,21 @@ module Haml
         end
 
         output << ">" if nuke_outer_whitespace
-        output << "/" if empty? && !etag
+        output << "/" if to_xhtml.end_with?("/>")
 
         if children && children.size == 1
           child = children.first
-          if child.is_a?(::Hpricot::Text)
+          if child.is_a?(::Nokogiri::XML::Text)
             if !child.to_s.include?("\n")
               text = child.to_haml(tabs + 1, options)
               return output + " " + text.lstrip.gsub(/^\\/, '') unless text.chomp.include?("\n") || text.empty?
               return output + "\n" + text
             elsif ["pre", "textarea"].include?(name) ||
-                (name == "code" && parent.is_a?(::Hpricot::Elem) && parent.name == "pre")
+                (name == "code" && parent.is_a?(::Nokogiri::XML::Element) && parent.name == "pre")
               return output + "\n#{tabulate(tabs + 1)}:preserve\n" +
-                innerText.gsub(/^/, tabulate(tabs + 2))
+                inner_text.gsub(/^/, tabulate(tabs + 2))
             end
-          elsif child.is_a?(::Hpricot::Elem) && child.name == "haml:loud"
+          elsif child.is_a?(::Nokogiri::XML::Element) && child.name == "haml_loud"
             return output + child.to_haml(tabs + 1, options).lstrip
           end
         end
@@ -344,24 +387,30 @@ module Haml
       def dynamic_attributes
         @dynamic_attributes ||= begin
           Hash[attr_hash.map do |name, value|
-            next if value.empty?
-            full_match = nil
-            ruby_value = value.gsub(%r{<haml:loud>\s*(.+?)\s*</haml:loud>}) do
-              full_match = $`.empty? && $'.empty?
-              CGI.unescapeHTML(full_match ? $1: "\#{#{$1}}")
+            if value == ""
+              [nil, nil]
+            else
+              full_match = nil
+              ruby_value = value.to_s.gsub(%r{<haml_loud>\s*(.+?)\s*</haml_loud>}) do
+                full_match = $`.empty? && $'.empty?
+                CGI.unescapeHTML(full_match ? $1: "\#{#{$1}}")
+              end
+              if ruby_value == value
+                [nil, nil]
+              else
+                [name.to_s, full_match ? ruby_value : %("#{ruby_value}")]
+              end
             end
-            next if ruby_value == value
-            [name, full_match ? ruby_value : %("#{ruby_value}")]
           end]
         end
       end
 
       def to_haml_filter(filter, tabs, options)
         content =
-          if children.first.is_a?(::Hpricot::CData)
-            children.first.content
+          if children.first.cdata?
+            decode_entities(children.first.content_without_cdata_tokens)
           else
-            CGI.unescapeHTML(self.innerText)
+            decode_entities(self.inner_text)
           end
 
         content = erb_to_interpolation(content, options)
@@ -382,6 +431,17 @@ module Haml
         "#{tabulate(tabs)}:#{filter}\n#{content}"
       end
 
+      # TODO: this method is utterly awful, find a better way to decode HTML entities.
+      def decode_entities(str)
+        str.gsub(/&[\S]+;/) do |entity|
+          begin
+            [Nokogiri::HTML::NamedCharacters[entity[1..-2]]].pack("C")
+          rescue TypeError
+            entity
+          end
+        end
+      end
+
       def static_attribute?(name, options)
         attr_hash[name] && !dynamic_attribute?(name, options)
       end
@@ -406,7 +466,7 @@ module Haml
       # that's prettier than that produced by Hash#inspect
       def haml_attributes(options)
         attrs = attr_hash.sort.map do |name, value|
-          haml_attribute_pair(name, value, options)
+          haml_attribute_pair(name, value.to_s, options)
         end
         if options[:html_style_attributes]
           "(#{attrs.join(' ')})"
@@ -418,12 +478,20 @@ module Haml
       # Returns the string representation of a single attribute key value pair
       def haml_attribute_pair(name, value, options)
         value = dynamic_attribute?(name, options) ? dynamic_attributes[name] : value.inspect
+
         if options[:html_style_attributes]
-          "#{name}=#{value}"
-        else
-          name = name.index(/\W/) ? name.inspect : ":#{name}"
-          "#{name} => #{value}"
+          return "#{name}=#{value}"
+        end
+
+        if name.index(/\W/)
+          return "#{name.inspect} => #{value}"
         end
+
+        if options[:ruby19_style_attributes]
+          return "#{name}: #{value}"
+        end
+
+        ":#{name} => #{value}"
       end
     end
   end

data/lib/html2haml/html/erb.rb

@@ -2,10 +2,10 @@ require 'cgi'
 require 'erubis'
 require 'ruby_parser'
 
-module Haml
+module Html2haml
   class HTML
     # A class for converting ERB code into a format that's easier
-    # for the {Haml::HTML} Hpricot-based parser to understand.
+    # for the {Html2haml::HTML} Nokogiri-based parser to understand.
     #
     # Uses [Erubis](http://www.kuwata-lab.com/erubis)'s extensible parsing powers
     # to parse the ERB in a reliable way,
@@ -13,17 +13,17 @@ module Haml
     # to figure out whether a given chunk of Ruby code starts a block or not.
     #
     # The ERB tags are converted to HTML tags in the following way.
-    # `<% ... %>` is converted into `<haml:silent> ... </haml:silent>`.
-    # `<%= ... %>` is converted into `<haml:loud> ... </haml:loud>`.
+    # `<% ... %>` is converted into `<haml_silent> ... </haml_silent>`.
+    # `<%= ... %>` is converted into `<haml_loud> ... </haml_loud>`.
     # Finally, if either of these opens a Ruby block,
-    # `<haml:block> ... </haml:block>` will wrap the entire contents of the block -
+    # `<haml_block> ... </haml_block>` will wrap the entire contents of the block -
     # that is, everything that should be indented beneath the previous silent or loud tag.
     class ERB < Erubis::Basic::Engine
-      # Compiles an ERB template into a HTML document containing `haml:` tags.
+      # Compiles an ERB template into a HTML document containing `haml_*` tags.
       #
       # @param template [String] The ERB template
       # @return [String] The output document
-      # @see Haml::HTML::ERB
+      # @see Html2haml::HTML::ERB
       def self.compile(template)
         new(template).src
       end
@@ -48,8 +48,8 @@ module Haml
       end
 
       # Concatenates a silent Ruby statement onto the source buffer.
-      # This uses the `<haml:silent>` tag,
-      # and may close and/or open a Ruby block with the `<haml:block>` tag.
+      # This uses the `<haml_silent>` tag,
+      # and may close and/or open a Ruby block with the `<haml_block>` tag.
       #
       # In particular, a block is closed if this statement is some form of `end`,
       # opened if it's a block opener like `do`, `if`, or `begin`,
@@ -59,9 +59,9 @@ module Haml
       # @param src [String] The source buffer
       # @param code [String] The Ruby statement to add to the buffer
       def add_stmt(src, code)
-        src << '</haml:block>' if block_closer?(code) || mid_block?(code)
-        src << '<haml:silent>' << h(code) << '</haml:silent>' unless code.strip == "end"
-        src << '<haml:block>' if block_opener?(code) || mid_block?(code)
+        src << '</haml_block>' if block_closer?(code) || mid_block?(code)
+        src << '<haml_silent>' << h(code) << '</haml_silent>' unless code.strip == "end"
+        src << '<haml_block>' if block_opener?(code) || mid_block?(code)
       end
 
       # Concatenates a Ruby expression that's printed to the document
@@ -73,8 +73,8 @@ module Haml
       # @param src [String] The source buffer
       # @param code [String] The Ruby expression to add to the buffer
       def add_expr_literal(src, code)
-        src << '<haml:loud>' << h(code) << '</haml:loud>'
-        src << '<haml:block>' if block_opener?(code)
+        src << '<haml_loud>' << h(code) << '</haml_loud>'
+        src << '<haml_block>' if block_opener?(code)
       end
 
       # `html2haml` doesn't support debugging expressions.

data/lib/html2haml/version.rb

@@ -1,3 +1,3 @@
 module Html2haml
-  VERSION = "1.0.1"
+  VERSION = "2.0.0.beta.1"
 end

data/test/erb_test.rb

@@ -10,6 +10,12 @@ class ErbTest < MiniTest::Unit::TestCase
 
   def test_inline_erb
     assert_equal("%p= foo", render_erb("<p><%= foo %></p>"))
+    assert_equal(<<HAML.rstrip, render_erb(<<HTML))
+%p= foo
+HAML
+<p><%= foo %>
+</p>
+HTML
   end
 
   def test_non_inline_erb
@@ -27,13 +33,6 @@ HTML
 HAML
 <p>
   <%= foo %></p>
-HTML
-    assert_equal(<<HAML.rstrip, render_erb(<<HTML))
-%p
-  = foo
-HAML
-<p><%= foo %>
-</p>
 HTML
   end
 
@@ -460,15 +459,15 @@ ERB
   def test_can_parse_ruby_19_hashes_as_arguments
     erb =  "<%= foobar 'foo', {bar: 'baz'} %>"
     begin
-      Haml::HTML::ERB.new(erb)
+      Html2haml::HTML::ERB.new(erb)
     rescue
       flunk "should not raise an error"
     end
   end
 
   def test_should_wrap_in_silent
-    assert_equal(<<HTML.rstrip, Haml::HTML::ERB.new(<<ERB).src)
-<haml:silent> some_variable_or_function \n</haml:silent>
+    assert_equal(<<HTML.rstrip, Html2haml::HTML::ERB.new(<<ERB).src)
+<haml_silent> some_variable_or_function \n</haml_silent>
 HTML
 <% some_variable_or_function %>
 ERB
@@ -476,8 +475,8 @@ ERB
 
   #comment content is removed by erubis
   def test_should_wrap_process_comments_as_empty_lines
-    assert_equal(<<HTML.rstrip, Haml::HTML::ERB.new(<<ERB).src)
-<haml:silent>\n</haml:silent>
+    assert_equal(<<HTML.rstrip, Html2haml::HTML::ERB.new(<<ERB).src)
+<haml_silent>\n</haml_silent>
 HTML
 <%# some_variable_or_function %>
 ERB

data/test/html2haml_test.rb

@@ -45,6 +45,13 @@ class Html2HamlTest < MiniTest::Unit::TestCase
       render('<meta http-equiv="Content-Type" content="text/html" />', :html_style_attributes => true))
   end
 
+  def test_should_have_ruby_19_hash_style_attributes
+    assert_equal('%input{name: "login", type: "text"}/',
+      render('<input type="text" name="login" />', :ruby19_style_attributes => true))
+    assert_equal('%meta{content: "text/html", "http-equiv" => "Content-Type"}/',
+      render('<meta http-equiv="Content-Type" content="text/html" />', :ruby19_style_attributes => true))
+  end
+
   def test_class_with_dot_and_hash
     assert_equal('%div{:class => "foo.bar"}', render("<div class='foo.bar'></div>"))
     assert_equal('%div{:class => "foo#bar"}', render("<div class='foo#bar'></div>"))
@@ -81,7 +88,7 @@ HTML
   end
 
   def test_self_closing_tag
-    assert_equal("%foo/", render("<foo />"))
+    assert_equal("%img/", render("<img />"))
   end
 
   def test_inline_text
@@ -142,6 +149,21 @@ HTML
       return "12" & "13";
   }
 HAML
+<script type="text/javascript">
+    function foo() {
+        return "12" & "13";
+    }
+</script>
+HTML
+  end
+
+  def test_script_tag_with_html_escaped_javascript
+    assert_equal(<<HAML.rstrip, render(<<HTML))
+:javascript
+  function foo() {
+      return "12" & "13";
+  }
+HAML
 <script type="text/javascript">
     function foo() {
         return "12" &amp; "13";
@@ -333,24 +355,22 @@ HTML
 
   # Encodings
 
-  unless RUBY_VERSION < "1.9"
-    def test_encoding_error
-      render("foo\nbar\nb\xFEaz".force_encoding("utf-8"))
-      assert(false, "Expected exception")
-    rescue Haml::Error => e
-      assert_equal(3, e.line)
-      assert_equal('Invalid UTF-8 character "\xFE"', e.message)
-    end
-
-    def test_ascii_incompatible_encoding_error
-      template = "foo\nbar\nb_z".encode("utf-16le")
-      template[9] = "\xFE".force_encoding("utf-16le")
-      render(template)
-      assert(false, "Expected exception")
-    rescue Haml::Error => e
-      assert_equal(3, e.line)
-      assert_equal('Invalid UTF-16LE character "\xFE"', e.message)
-    end
+  def test_encoding_error
+    render("foo\nbar\nb\xFEaz".force_encoding("utf-8"))
+    assert(false, "Expected exception")
+  rescue Haml::Error => e
+    assert_equal(3, e.line)
+    assert_match(/Invalid UTF-8 character/, e.message)
+  end
+
+  def test_ascii_incompatible_encoding_error
+    template = "foo\nbar\nb_z".encode("utf-16le")
+    template[9] = "\xFE".force_encoding("utf-16le")
+    render(template)
+    assert(false, "Expected exception")
+  rescue Haml::Error => e
+    assert_equal(3, e.line)
+    assert_match(/Invalid UTF-16LE character/, e.message)
   end
 
   # Regression Tests
@@ -361,4 +381,25 @@ HTML
   "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
 HTML
   end
+
+  def test_html_document_without_doctype
+    assert_equal(<<HAML.rstrip, render(<<HTML))
+!!!
+%html
+  %head
+    %title Hello
+  %body
+    %p Hello
+HAML
+<html>
+<head>
+  <title>Hello</title>
+</head>
+<body>
+  <p>Hello</p>
+</body>
+</html>
+HTML
+  end
+
 end

data/test/test_helper.rb

@@ -13,7 +13,7 @@ require 'html2haml/html/erb'
 class MiniTest::Unit::TestCase
   protected
   def render(text, options = {})
-    Haml::HTML.new(text, options).render.rstrip
+    Html2haml::HTML.new(text, options).render.rstrip
   end
 
   def render_erb(text)

metadata

@@ -1,132 +1,126 @@
---- !ruby/object:Gem::Specification 
+--- !ruby/object:Gem::Specification
 name: html2haml
-version: !ruby/object:Gem::Version 
-  prerelease: false
-  segments: 
-  - 1
-  - 0
-  - 1
-  version: 1.0.1
+version: !ruby/object:Gem::Version
+  version: 2.0.0.beta.1
 platform: ruby
-authors: 
+authors:
 - Norman Clarke
+- Stefan Natchev
 autorequire: 
 bindir: bin
 cert_chain: []
-
-date: 2013-02-16 00:00:00 -03:00
-default_executable: 
-dependencies: 
-- !ruby/object:Gem::Dependency 
-  name: hpricot
-  prerelease: false
-  requirement: &id001 !ruby/object:Gem::Requirement 
-    requirements: 
+date: 2013-07-15 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: nokogiri
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 0
-        - 8
-        - 6
-        version: 0.8.6
+      - !ruby/object:Gem::Version
+        version: 1.6.0
   type: :runtime
-  version_requirements: *id001
-- !ruby/object:Gem::Dependency 
-  name: erubis
   prerelease: false
-  requirement: &id002 !ruby/object:Gem::Requirement 
-    requirements: 
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 2
-        - 7
-        - 0
+      - !ruby/object:Gem::Version
+        version: 1.6.0
+- !ruby/object:Gem::Dependency
+  name: erubis
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
         version: 2.7.0
   type: :runtime
-  version_requirements: *id002
-- !ruby/object:Gem::Dependency 
-  name: ruby_parser
   prerelease: false
-  requirement: &id003 !ruby/object:Gem::Requirement 
-    requirements: 
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 3
-        - 1
-        - 1
-        version: 3.1.1
+      - !ruby/object:Gem::Version
+        version: 2.7.0
+- !ruby/object:Gem::Dependency
+  name: ruby_parser
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 3.2.1
   type: :runtime
-  version_requirements: *id003
-- !ruby/object:Gem::Dependency 
-  name: haml
   prerelease: false
-  requirement: &id004 !ruby/object:Gem::Requirement 
-    requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 4
-        - 0
-        - 0
-        - rc
-        - 1
-        version: 4.0.0.rc.1
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 3.2.1
+- !ruby/object:Gem::Dependency
+  name: haml
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 4.0.0
   type: :runtime
-  version_requirements: *id004
-- !ruby/object:Gem::Dependency 
-  name: simplecov
   prerelease: false
-  requirement: &id005 !ruby/object:Gem::Requirement 
-    requirements: 
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 4.0.0
+- !ruby/object:Gem::Dependency
+  name: simplecov
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 0
-        - 7
-        - 1
+      - !ruby/object:Gem::Version
         version: 0.7.1
   type: :development
-  version_requirements: *id005
-- !ruby/object:Gem::Dependency 
-  name: minitest
   prerelease: false
-  requirement: &id006 !ruby/object:Gem::Requirement 
-    requirements: 
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 0.7.1
+- !ruby/object:Gem::Dependency
+  name: minitest
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 4
-        - 4
-        - 0
+      - !ruby/object:Gem::Version
         version: 4.4.0
   type: :development
-  version_requirements: *id006
-- !ruby/object:Gem::Dependency 
-  name: rake
   prerelease: false
-  requirement: &id007 !ruby/object:Gem::Requirement 
-    requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 0
-        version: "0"
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 4.4.0
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
   type: :development
-  version_requirements: *id007
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
 description: Converts HTML into Haml
-email: 
+email:
 - norman@njclarke.com
-executables: 
+- stefan.natchev@gmail.com
+executables:
 - html2haml
 extensions: []
-
 extra_rdoc_files: []
-
-files: 
+files:
 - .gitignore
 - .travis.yml
+- .yardopts
 - Changelog.markdown
 - Gemfile
 - MIT-LICENSE
@@ -142,37 +136,30 @@ files:
 - test/erb_test.rb
 - test/html2haml_test.rb
 - test/test_helper.rb
-has_rdoc: true
 homepage: http://haml.info
 licenses: []
-
+metadata: {}
 post_install_message: 
 rdoc_options: []
-
-require_paths: 
+require_paths:
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement 
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      segments: 
-      - 0
-      version: "0"
-required_rubygems_version: !ruby/object:Gem::Requirement 
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      segments: 
-      - 0
-      version: "0"
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: 1.9.2
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>'
+    - !ruby/object:Gem::Version
+      version: 1.3.1
 requirements: []
-
 rubyforge_project: 
-rubygems_version: 1.3.6
+rubygems_version: 2.0.3
 signing_key: 
-specification_version: 3
+specification_version: 4
 summary: Converts HTML into Haml
-test_files: 
+test_files:
 - test/erb_test.rb
 - test/html2haml_test.rb
 - test/test_helper.rb