Merged pull request #181 from jkingdon/nokogiri-wsdl-parsing.

For parsing WSDL, switch from REXML to Nokogiri and stream parsing to DOM

lib/savon/wsdl/document.rb

@@ -1,4 +1,4 @@
-require "rexml/document"
+require "nokogiri"
 
 require "savon/wsdl/request"
 require "savon/wsdl/parser"
@@ -101,8 +101,8 @@ def read_file
       # Parses the WSDL document and returns the <tt>Savon::WSDL::Parser</tt>.
       def parser
         @parser ||= begin
-          parser = Parser.new
-          REXML::Document.parse_stream document, parser
+          parser = Parser.new(Nokogiri::XML(document))
+          parser.parse
           parser
         end
       end

lib/savon/wsdl/parser.rb

@@ -6,13 +6,14 @@ module WSDL
 
     # = Savon::WSDL::Parser
     #
-    # Serves as a stream listener for parsing WSDL documents.
+    # Parses WSDL documents and remembers the parts of them we care about.
     class Parser
 
       # The main sections of a WSDL document.
       Sections = %w(definitions types message portType binding service)
 
-      def initialize
+      def initialize(nokogiri_document)
+        @document = nokogiri_document
         @path = []
         @operations = {}
         @namespaces = {}
@@ -31,72 +32,65 @@ def initialize
       # Returns the elementFormDefault value.
       attr_reader :element_form_default
 
-      # Hook method called when the stream parser encounters a starting tag.
-      def tag_start(tag, attrs)
-        # read xml namespaces if root element
-        read_namespaces(attrs) if @path.empty?
-
-        tag, namespace = tag.split(":").reverse
-        @path << tag
-
-        if @section == :types && tag == "schema"
-          @element_form_default = attrs["elementFormDefault"].to_sym if attrs["elementFormDefault"]
-        end
-
-        if @section == :binding && tag == "binding"
-          # ensure that we are in an wsdl/soap namespace
-          @section = nil unless @namespaces[namespace].starts_with? "http://schemas.xmlsoap.org/wsdl/soap"
-        end
-
-        @section = tag.to_sym if Sections.include?(tag) && depth <= 2
-
-        @namespace ||= attrs["targetNamespace"] if @section == :definitions
-        @endpoint ||= URI(URI.escape(attrs["location"])) if @section == :service && tag == "address"
-
-        operation_from tag, attrs if @section == :binding && tag == "operation"
-      end
-
-      # Returns our current depth in the WSDL document.
-      def depth
-        @path.size
+      def parse
+        parse_namespaces
+        parse_endpoint
+        parse_operations
       end
 
-      # Reads namespace definitions from a given +attrs+ Hash.
-      def read_namespaces(attrs)
-        attrs.each do |key, value|
-          @namespaces[key.strip_namespace] = value if key.starts_with? "xmlns:"
-        end
+      def parse_namespaces
+        element_form_default = @document.at_xpath(
+          "s0:definitions/s0:types/xs:schema/@elementFormDefault",
+          "s0" => "http://schemas.xmlsoap.org/wsdl/",
+          "xs" => "http://www.w3.org/2001/XMLSchema")
+        @element_form_default = element_form_default.to_s.to_sym if element_form_default
+
+        namespace = @document.at_xpath(
+          "s0:definitions/@targetNamespace",
+          "s0" => "http://schemas.xmlsoap.org/wsdl/")
+        @namespace = namespace.to_s if namespace
       end
 
-      # Hook method called when the stream parser encounters a closing tag.
-      def tag_end(tag)
-        @path.pop
-
-        if @section == :binding && @input && tag.strip_namespace == "operation"
-          # no soapAction attribute found till now
-          operation_from tag, "soapAction" => @input
-        end
-
-        @section = :definitions if Sections.include?(tag) && depth <= 1
+      def parse_endpoint
+        endpoint = @document.at_xpath(
+          "s0:definitions/s0:service//soap11:address/@location",
+          "s0" => "http://schemas.xmlsoap.org/wsdl/",
+          "soap11" => "http://schemas.xmlsoap.org/wsdl/soap/")
+        endpoint ||= @document.at_xpath(
+          "s0:definitions/s0:service//soap12:address/@location",
+          "s0" => "http://schemas.xmlsoap.org/wsdl/",
+          "soap12" => "http://schemas.xmlsoap.org/wsdl/soap12/")
+        @endpoint = URI(URI.escape(endpoint.to_s)) if endpoint
       end
 
-      # Stores available operations from a given tag +name+ and +attrs+.
-      def operation_from(tag, attrs)
-        @input = attrs["name"] if attrs["name"]
-
-        if attrs["soapAction"]
-          @action = !attrs["soapAction"].blank? ? attrs["soapAction"] : @input
-          @input = @action.split("/").last if !@input || @input.empty?
-
-          @operations[@input.snakecase.to_sym] = { :action => @action, :input => @input }
-          @input, @action = nil, nil
+      def parse_operations
+        operations = @document.xpath(
+          "s0:definitions/s0:binding/s0:operation",
+          "s0" => "http://schemas.xmlsoap.org/wsdl/")
+        operations.each do |operation|
+          name = operation.attribute("name").to_s
+
+          soap_action = operation.at_xpath(".//soap11:operation/@soapAction",
+            "soap11" => "http://schemas.xmlsoap.org/wsdl/soap/"
+          )
+          soap_action ||= operation.at_xpath(".//soap12:operation/@soapAction",
+            "soap12" => "http://schemas.xmlsoap.org/wsdl/soap12/"
+          )
+          if soap_action
+            soap_action = soap_action.to_s
+
+            action = !soap_action.blank? ? soap_action : name
+            input = (!name || name.empty?) ? action.split("/").last : name
+
+            @operations[input.snakecase.to_sym] =
+              { :action => action, :input => input }
+          elsif !@operations[name.snakecase.to_sym]
+            @operations[name.snakecase.to_sym] =
+              { :action => name, :input => name }
+          end
         end
       end
 
-      # Catches calls to unimplemented hook methods.
-      def method_missing(method, *args)
-      end
-
     end
   end
 end

savon.gemspec

@@ -18,6 +18,7 @@ Gem::Specification.new do |s|
   s.add_dependency "crack", "~> 0.1.8"
   s.add_dependency "httpi", ">= 0.7.8"
   s.add_dependency "gyoku", ">= 0.4.0"
+  s.add_dependency "nokogiri"
 
   s.add_development_dependency "rspec", "~> 2.4.0"
   s.add_development_dependency "autotest"

spec/fixtures/wsdl/soap12.xml

@@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Example of a WSDL with SOAP 1.2 and no SOAP 1.1 endpoint.
+     Don't know whether this is widespread, but we should allow it. -->
+<definitions xmlns="http://schemas.xmlsoap.org/wsdl/" xmlns:soap="http://schemas.xmlsoap.org/wsdl/soap/" xmlns:soap12="http://schemas.xmlsoap.org/wsdl/soap12/" xmlns:wsdl="http://schemas.xmlsoap.org/wsdl/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
+    <service name="Blog">
+        <port name="BlogSoap12">
+            <soap12:address location="http://blogsite.example.com/endpoint12"/>
+        </port>
+    </service>
+</definitions>
+

spec/savon/wsdl/parser_spec.rb

@@ -82,6 +82,14 @@
     end
   end
 
+  context "with soap12.xml" do
+    let(:parser) { new_parser :soap12 }
+
+    it "should return the endpoint" do
+      parser.endpoint.should == URI("http://blogsite.example.com/endpoint12")
+    end
+  end
+
   RSpec::Matchers.define :match_operations do |expected|
     match do |actual|
       actual.should have(expected.keys.size).items
@@ -91,8 +99,8 @@
   end
 
   def new_parser(fixture)
-    parser = Savon::WSDL::Parser.new
-    REXML::Document.parse_stream Fixture[:wsdl, fixture], parser
+    parser = Savon::WSDL::Parser.new(Nokogiri::XML(Fixture[:wsdl, fixture]))
+    parser.parse
     parser
   end