Added Hash.create_from_xml(string) which will create a hash from a XML string and even typecast if possible [DHH]

git-svn-id: http://svn-commit.rubyonrails.org/rails/trunk@4453 5ecf4fe2-1ee6-0310-87b1-e25e094e27de

7 files changed, 236 insertions, 143 deletions

diff --git a/actionpack/lib/action_controller/cgi_ext/cgi_methods.rb b/actionpack/lib/action_controller/cgi_ext/cgi_methods.rb
index be5923e8de..1af837c39a 100755
--- a/actionpack/lib/action_controller/cgi_ext/cgi_methods.rb
+++ b/actionpack/lib/action_controller/cgi_ext/cgi_methods.rb
@@ -1,14 +1,13 @@
 require 'cgi'
-require 'action_controller/vendor/xml_simple'
 require 'action_controller/vendor/xml_node'
 
 # Static methods for parsing the query and request parameters that can be used in
 # a CGI extension class or testing in isolation.
 class CGIMethods #:nodoc:
-  public
+  class << self
     # Returns a hash with the pairs from the query string. The implicit hash construction that is done in
     # parse_request_params is not done here.
-    def CGIMethods.parse_query_parameters(query_string)
+    def parse_query_parameters(query_string)
       parsed_params = {}
   
       query_string.split(/[&;]/).each { |p| 
@@ -41,7 +40,7 @@ class CGIMethods #:nodoc:
     # Returns the request (POST/GET) parameters in a parsed form where pairs such as "customer[address][street]" / 
     # "Somewhere cool!" are translated into a full hash hierarchy, like
     # { "customer" => { "address" => { "street" => "Somewhere cool!" } } }
-    def CGIMethods.parse_request_parameters(params)
+    def parse_request_parameters(params)
       parsed_params = {}
 
       for key, value in params
@@ -59,162 +58,103 @@ class CGIMethods #:nodoc:
       parsed_params
     end
 
-    def self.parse_formatted_request_parameters(mime_type, raw_post_data)
-      params = case strategy = ActionController::Base.param_parsers[mime_type]
+    def parse_formatted_request_parameters(mime_type, raw_post_data)
+      case strategy = ActionController::Base.param_parsers[mime_type]
         when Proc
           strategy.call(raw_post_data)
         when :xml_simple
-          raw_post_data.blank? ? nil :
-            typecast_xml_value(XmlSimple.xml_in(raw_post_data,
-              'forcearray'   => false,
-              'forcecontent' => true,
-              'keeproot'     => true,
-              'contentkey'   => '__content__'))
+          raw_post_data.blank? ? {} : Hash.create_from_xml(raw_post_data)
         when :yaml
           YAML.load(raw_post_data)
         when :xml_node
           node = XmlNode.from_xml(raw_post_data)
           { node.node_name => node }
       end
-      
-      dasherize_keys(params || {})
     rescue Object => e
       { "exception" => "#{e.message} (#{e.class})", "backtrace" => e.backtrace, 
         "raw_post_data" => raw_post_data, "format" => mime_type }
     end
 
-    def self.typecast_xml_value(value)
-      case value
-      when Hash
-        if value.has_key?("__content__")
-          content = translate_xml_entities(value["__content__"])
-          case value["type"]
-          when "integer"  then content.to_i
-          when "boolean"  then content == "true"
-          when "datetime" then Time.parse(content)
-          when "date"     then Date.parse(content)
-          else                 content
-          end
-        else
-          value.empty? ? nil : value.inject({}) do |h,(k,v)|
-            h[k] = typecast_xml_value(v)
-            h
-          end
-        end
-      when Array
-        value.map! { |i| typecast_xml_value(i) }
-        case value.length
-        when 0 then nil
-        when 1 then value.first
-        else value
-        end
-      else
-        raise "can't typecast #{value.inspect}"
-      end
-    end
-
-  private
-
-    def self.translate_xml_entities(value)
-      value.gsub(/&lt;/,   "<").
-            gsub(/&gt;/,   ">").
-            gsub(/&quot;/, '"').
-            gsub(/&apos;/, "'").
-            gsub(/&amp;/,  "&")
-    end
-
-    def self.dasherize_keys(params)
-      case params.class.to_s
-      when "Hash"
-        params.inject({}) do |h,(k,v)|
-          h[k.to_s.tr("-", "_")] = dasherize_keys(v)
-          h
-        end
-      when "Array"
-        params.map { |v| dasherize_keys(v) }
-      else
-        params
-      end
-    end
-
-    # Splits the given key into several pieces. Example keys are 'name', 'person[name]',
-    # 'person[name][first]', and 'people[]'. In each instance, an Array instance is returned.
-    # 'person[name][first]' produces ['person', 'name', 'first']; 'people[]' produces ['people', '']
-    def CGIMethods.split_key(key)
-      if /^([^\[]+)((?:\[[^\]]*\])+)$/ =~ key
-        keys = [$1]
+    private
+      # Splits the given key into several pieces. Example keys are 'name', 'person[name]',
+      # 'person[name][first]', and 'people[]'. In each instance, an Array instance is returned.
+      # 'person[name][first]' produces ['person', 'name', 'first']; 'people[]' produces ['people', '']
+      def split_key(key)
+        if /^([^\[]+)((?:\[[^\]]*\])+)$/ =~ key
+          keys = [$1]
         
-        keys.concat($2[1..-2].split(']['))
-        keys << '' if key[-2..-1] == '[]' # Have to add it since split will drop empty strings
+          keys.concat($2[1..-2].split(']['))
+          keys << '' if key[-2..-1] == '[]' # Have to add it since split will drop empty strings
         
-        keys
-      else
-        [key]
+          keys
+        else
+          [key]
+        end
       end
-    end
     
-    def CGIMethods.get_typed_value(value)
-      # test most frequent case first
-      if value.is_a?(String)
-        value
-      elsif value.respond_to?(:content_type) && ! value.content_type.blank?
-        # Uploaded file
-        unless value.respond_to?(:full_original_filename)
-          class << value
-            alias_method :full_original_filename, :original_filename
-
-            # Take the basename of the upload's original filename.
-            # This handles the full Windows paths given by Internet Explorer
-            # (and perhaps other broken user agents) without affecting
-            # those which give the lone filename.
-            # The Windows regexp is adapted from Perl's File::Basename.
-            def original_filename
-              if md = /^(?:.*[:\\\/])?(.*)/m.match(full_original_filename)
-                md.captures.first
-              else
-                File.basename full_original_filename
+      def get_typed_value(value)
+        # test most frequent case first
+        if value.is_a?(String)
+          value
+        elsif value.respond_to?(:content_type) && ! value.content_type.blank?
+          # Uploaded file
+          unless value.respond_to?(:full_original_filename)
+            class << value
+              alias_method :full_original_filename, :original_filename
+
+              # Take the basename of the upload's original filename.
+              # This handles the full Windows paths given by Internet Explorer
+              # (and perhaps other broken user agents) without affecting
+              # those which give the lone filename.
+              # The Windows regexp is adapted from Perl's File::Basename.
+              def original_filename
+                if md = /^(?:.*[:\\\/])?(.*)/m.match(full_original_filename)
+                  md.captures.first
+                else
+                  File.basename full_original_filename
+                end
               end
             end
           end
-        end
 
-        # Return the same value after overriding original_filename.
-        value
+          # Return the same value after overriding original_filename.
+          value
 
-      elsif value.respond_to?(:read)
-        # Value as part of a multipart request
-        result = value.read
-        value.rewind
-        result
-      elsif value.class == Array
-        value.collect { |v| CGIMethods.get_typed_value(v) }
-      else
-        # other value (neither string nor a multipart request)
-        value.to_s
+        elsif value.respond_to?(:read)
+          # Value as part of a multipart request
+          result = value.read
+          value.rewind
+          result
+        elsif value.class == Array
+          value.collect { |v| get_typed_value(v) }
+        else
+          # other value (neither string nor a multipart request)
+          value.to_s
+        end
       end
-    end
   
-    PARAMS_HASH_RE = /^([^\[]+)(\[.*\])?(.)?.*$/
-    def CGIMethods.get_levels(key)
-      all, main, bracketed, trailing = PARAMS_HASH_RE.match(key).to_a
-      if main.nil?
-        []
-      elsif trailing
-        [key]
-      elsif bracketed
-        [main] + bracketed.slice(1...-1).split('][')
-      else
-        [main]
+      PARAMS_HASH_RE = /^([^\[]+)(\[.*\])?(.)?.*$/
+      def get_levels(key)
+        all, main, bracketed, trailing = PARAMS_HASH_RE.match(key).to_a
+        if main.nil?
+          []
+        elsif trailing
+          [key]
+        elsif bracketed
+          [main] + bracketed.slice(1...-1).split('][')
+        else
+          [main]
+        end
       end
-    end
 
-    def CGIMethods.build_deep_hash(value, hash, levels)
-      if levels.length == 0
-        value
-      elsif hash.nil?
-        { levels.first => CGIMethods.build_deep_hash(value, nil, levels[1..-1]) }
-      else
-        hash.update({ levels.first => CGIMethods.build_deep_hash(value, hash[levels.first], levels[1..-1]) })
+      def build_deep_hash(value, hash, levels)
+        if levels.length == 0
+          value
+        elsif hash.nil?
+          { levels.first => build_deep_hash(value, nil, levels[1..-1]) }
+        else
+          hash.update({ levels.first => build_deep_hash(value, hash[levels.first], levels[1..-1]) })
+        end
       end
-    end
+  end
 end
diff --git a/actionpack/test/controller/webservice_test.rb b/actionpack/test/controller/webservice_test.rb
index 9874b2092b..e60e3a84cb 100644
--- a/actionpack/test/controller/webservice_test.rb
+++ b/actionpack/test/controller/webservice_test.rb
@@ -148,13 +148,6 @@ class WebServiceTest < Test::Unit::TestCase
     assert_equal %(<foo "bar's" & friends>), @controller.params[:data]
   end
 
-  def test_dasherized_keys_as_yaml
-    ActionController::Base.param_parsers[Mime::YAML] = :yaml
-    process('POST', 'application/x-yaml', "---\nfirst-key:\n  sub-key: ...\n", true)
-    assert_equal 'action, controller, first_key(sub_key), full', @controller.response.body
-    assert_equal "...", @controller.params[:first_key][:sub_key]
-  end
-
   def test_typecast_as_yaml
     ActionController::Base.param_parsers[Mime::YAML] = :yaml
     process('POST', 'application/x-yaml', <<-YAML)
diff --git a/activesupport/CHANGELOG b/activesupport/CHANGELOG
index 9394fcfed5..5a5fc9d98e 100644
--- a/activesupport/CHANGELOG
+++ b/activesupport/CHANGELOG
@@ -1,5 +1,18 @@
 *SVN*
 
+* Added Hash.create_from_xml(string) which will create a hash from a XML string and even typecast if possible [DHH]. Example:
+
+    Hash.create_from_xml <<-EOT
+      <note>
+        <title>This is a note</title>
+        <created-at type="date">2004-10-10</created-at>
+      </note>
+    EOT
+  
+  ...would return:
+  
+    { :note => { :title => "This is a note", :created_at => Date.new(2004, 10, 10) } }
+
 * Added Jim Weirich's excellent FlexMock class to vendor (Copyright 2003, 2004 by Jim Weirich (jim@weriichhouse.org)) -- it's not automatically required, though, so require 'flexmock' is still necessary [DHH]
 
 * Fixed that Module#alias_method_chain should work with both foo? foo! and foo at the same time #4954 [anna@wota.jp]
diff --git a/activesupport/lib/active_support/core_ext/hash/conversions.rb b/activesupport/lib/active_support/core_ext/hash/conversions.rb
index a3270c31ef..ac4a612e49 100644
--- a/activesupport/lib/active_support/core_ext/hash/conversions.rb
+++ b/activesupport/lib/active_support/core_ext/hash/conversions.rb
@@ -1,4 +1,5 @@
 require 'date'
+require 'xml_simple'
 
 module ActiveSupport #:nodoc:
   module CoreExtensions #:nodoc:
@@ -20,6 +21,10 @@ module ActiveSupport #:nodoc:
           "binary"   => Proc.new { |binary| Base64.encode64(binary) }
         }
 
+        def self.included(klass)
+          klass.extend(ClassMethods)
+        end
+
         def to_xml(options = {})
           options[:indent] ||= 2
           options.reverse_merge!({ :builder => Builder::XmlMarkup.new(:indent => options[:indent]),
@@ -70,6 +75,71 @@ module ActiveSupport #:nodoc:
           end
 
         end
+
+        module ClassMethods
+          def create_from_xml(xml)
+            # TODO: Refactor this into something much cleaner that doesn't rely on XmlSimple
+            undasherize_keys(typecast_xml_value(XmlSimple.xml_in(xml,
+              'forcearray'   => false,
+              'forcecontent' => true,
+              'keeproot'     => true,
+              'contentkey'   => '__content__')
+            ))
+          end
+
+          private
+            def typecast_xml_value(value)
+              case value.class.to_s
+                when "Hash"
+                  if value.has_key?("__content__")
+                    content = translate_xml_entities(value["__content__"])
+                    case value["type"]
+                      when "integer"  then content.to_i
+                      when "boolean"  then content == "true"
+                      when "datetime" then ::Time.parse(content).utc
+                      when "date"     then ::Date.parse(content)
+                      else                 content
+                    end
+                  else
+                    value.empty? ? nil : value.inject({}) do |h,(k,v)|
+                      h[k] = typecast_xml_value(v)
+                      h
+                    end
+                  end
+                when "Array"
+                  value.map! { |i| typecast_xml_value(i) }
+                  case value.length
+                    when 0 then nil
+                    when 1 then value.first
+                    else value
+                  end
+                else
+                  raise "can't typecast #{value.inspect}"
+              end
+            end
+
+            def translate_xml_entities(value)
+              value.gsub(/&lt;/,   "<").
+                    gsub(/&gt;/,   ">").
+                    gsub(/&quot;/, '"').
+                    gsub(/&apos;/, "'").
+                    gsub(/&amp;/,  "&")
+            end
+
+            def undasherize_keys(params)
+              case params.class.to_s
+                when "Hash"
+                  params.inject({}) do |h,(k,v)|
+                    h[k.to_s.tr("-", "_")] = undasherize_keys(v)
+                    h
+                  end
+                when "Array"
+                  params.map { |v| undasherize_keys(v) }
+                else
+                  params
+              end
+            end
+        end
       end
     end
   end
diff --git a/activesupport/lib/active_support/core_ext/string/inflections.rb b/activesupport/lib/active_support/core_ext/string/inflections.rb
index ebae190052..7107bc6945 100644
--- a/activesupport/lib/active_support/core_ext/string/inflections.rb
+++ b/activesupport/lib/active_support/core_ext/string/inflections.rb
@@ -1,4 +1,5 @@
 require File.dirname(__FILE__) + '/../../inflector' unless defined? Inflector
+
 module ActiveSupport #:nodoc:
   module CoreExtensions #:nodoc:
     module String #:nodoc:
@@ -6,7 +7,6 @@ module ActiveSupport #:nodoc:
       # For instance, you can figure out the name of a database from the name of a class.
       #   "ScaleScore".tableize => "scale_scores"
       module Inflections
-        
         # Returns the plural form of the word in the string.
         #
         # Examples
diff --git a/actionpack/lib/action_controller/vendor/xml_simple.rb b/activesupport/lib/active_support/vendor/xml_simple.rb
index e6ce63a839..e6ce63a839 100644
--- a/actionpack/lib/action_controller/vendor/xml_simple.rb
+++ b/activesupport/lib/active_support/vendor/xml_simple.rb
diff --git a/activesupport/test/core_ext/hash_ext_test.rb b/activesupport/test/core_ext/hash_ext_test.rb
index f8ed54c881..65050de15a 100644
--- a/activesupport/test/core_ext/hash_ext_test.rb
+++ b/activesupport/test/core_ext/hash_ext_test.rb
@@ -280,4 +280,81 @@ class HashToXmlTest < Test::Unit::TestCase
     assert xml.include?(%(<addresses><address><streets><street><name>))
   end
 
-end
+  def test_single_record_from_xml
+    topic_xml = <<-EOT
+      <topic>
+        <title>The First Topic</title>
+        <author-name>David</author-name>
+        <id type="integer">1</id>
+        <approved type="boolean">false</approved>
+        <replies-count type="integer">0</replies-count>
+        <written-on type="date">2003-07-16</written-on>
+        <viewed-at type="datetime">2003-07-16T09:28:00+0000</viewed-at>
+        <content>Have a nice day</content>
+        <author-email-address>david@loudthinking.com</author-email-address>
+        <parent-id></parent-id>
+      </topic>
+    EOT
+    
+    expected_topic_hash = {
+      :title => "The First Topic",
+      :author_name => "David",
+      :id => 1,
+      :approved => false,
+      :replies_count => 0,
+      :written_on => Date.new(2003, 7, 16),
+      :viewed_at => Time.utc(2003, 7, 16, 9, 28),
+      :content => "Have a nice day",
+      :author_email_address => "david@loudthinking.com",
+      :parent_id => nil
+    }.stringify_keys
+    
+    assert_equal expected_topic_hash, Hash.create_from_xml(topic_xml)["topic"]
+  end
+
+  def test_multiple_records_from_xml
+    topics_xml = <<-EOT
+      <topics>
+        <topic>
+          <title>The First Topic</title>
+          <author-name>David</author-name>
+          <id type="integer">1</id>
+          <approved type="boolean">false</approved>
+          <replies-count type="integer">0</replies-count>
+          <written-on type="date">2003-07-16</written-on>
+          <viewed-at type="datetime">2003-07-16T09:28:00+0000</viewed-at>
+          <content>Have a nice day</content>
+          <author-email-address>david@loudthinking.com</author-email-address>
+          <parent-id></parent-id>
+        </topic>
+        <topic>
+          <title>The Second Topic</title>
+          <author-name>Jason</author-name>
+          <id type="integer">1</id>
+          <approved type="boolean">false</approved>
+          <replies-count type="integer">0</replies-count>
+          <written-on type="date">2003-07-16</written-on>
+          <viewed-at type="datetime">2003-07-16T09:28:00+0000</viewed-at>
+          <content>Have a nice day</content>
+          <author-email-address>david@loudthinking.com</author-email-address>
+          <parent-id></parent-id>
+        </topic>
+      </topics>
+    EOT
+    
+    expected_topic_hash = {
+      :title => "The First Topic",
+      :author_name => "David",
+      :id => 1,
+      :approved => false,
+      :replies_count => 0,
+      :written_on => Date.new(2003, 7, 16),
+      :viewed_at => Time.utc(2003, 7, 16, 9, 28),
+      :content => "Have a nice day",
+      :author_email_address => "david@loudthinking.com",
+      :parent_id => nil
+    }.stringify_keys
+    
+    assert_equal expected_topic_hash, Hash.create_from_xml(topics_xml)["topics"]["topic"].first
+  end
+end
\ No newline at end of file