onix 0.5.1 → 0.6.0

data/CHANGELOG

@@ -1,3 +1,10 @@
+v0.6.0 (18th March 2009)
+- remove use of threads in ONIX::Reader
+  - a producer/consumer pattern was useful in the REXML stream parsing days, but
+    now LibXML's Reader binding provides a better alternative
+  - API left unchanged, this was all under the hood
+- bump required ROXML version to 2.5.2
+
 v0.5.1 (4th March 2009)
 - Fix a single letter typo
 

data/lib/onix.rb

@@ -3,7 +3,7 @@ require 'bigdecimal'
 require 'cgi'
 
 # ensure we load the correct gem versions
-gem 'roxml', '2.5.1'
+gem 'roxml', '2.5.2'
 gem 'andand'
 
 # and now load the actual gems
@@ -13,8 +13,8 @@ require 'andand'
 module ONIX
   module Version #:nodoc:
     Major = 0
-    Minor = 5
-    Tiny  = 1
+    Minor = 6
+    Tiny  = 0
 
     String = [Major, Minor, Tiny].join('.')
   end

data/lib/onix/reader.rb

@@ -1,5 +1,3 @@
-require 'thread'
-require 'timeout'
 require 'stringio'
 
 module ONIX
@@ -66,65 +64,56 @@ module ONIX
       end
 
       @product_klass = product_klass
+      @header = nil
 
-      # create a sized queue to store each product read from the file
-      # We use a separate thread to read products from the source file.
-      # This queue is a thread-safe way to transfer products from that
-      # thread back into the main one.
-      @queue = SizedQueue.new(100)
-
-      # launch a reader thread
-      Thread.abort_on_exception = true
-      Thread.new { read_input }
-
-      # don't return from the constructor until the reading thread
-      # has spun up and put at least one item into the queue. If
-      # it finds no Products in the file, it queues a nil, so we
-      # shouldn't get stuck here indefinitely
-      while @queue.size == 0
-        sleep 0.05
+      while @header.nil? 
+        obj = read_next
+        if obj.kind_of?(ONIX::Header)
+          @header = obj
+        end
       end
     end
 
     # Iterate over all the products in an ONIX file
     #
     def each(&block)
-      obj = @queue.pop
-      while !obj.nil?
+      while !(obj = read_next).nil?
         yield obj
-        obj = @queue.pop
       end
     end
 
     private
 
-    # Walk the ONIX file, and grab the bits we're interested in.
-    #
-    # High level attributes and the header are stored as attributes of the reader
-    # class. Products are placed in a queue, ready to be popped off when the
-    # user uses the each() method.
+    # Walk the ONIX file, and grab the next header or product fragment. If we
+    # encounter other useful bits of info along the way (encoding, etc) then
+    # store them for later.
     #
-    def read_input
+    def read_next
       while @reader.read
+
         @xml_lang    = @reader.xml_lang         if @xml_lang.nil?
         @xml_version = @reader.xml_version.to_f if @xml_version.nil?
         @encoding    = encoding_const_to_name(@reader.encoding) if @encoding.nil?
+
         if @reader.node_type == LibXML::XML::Reader::TYPE_DOCUMENT_TYPE
           uri = @reader.expand.to_s
           m, major, minor, rev = *uri.match(/.+(\d)\.(\d)\/(\d*).*/)
           @version = [major.to_i, minor.to_i, rev.to_i]
         elsif @reader.name == "Header" && @reader.node_type == LibXML::XML::Reader::TYPE_ELEMENT
-          @header = ONIX::Header.from_xml(@reader.expand.to_s)
+          str = @reader.read_outer_xml
           @reader.next_sibling
+          return ONIX::Header.from_xml(str)
         elsif @reader.name == "Product" && @reader.node_type == LibXML::XML::Reader::TYPE_ELEMENT
-          node = @reader.expand
-          @queue.push @product_klass.from_xml(node.to_s)
+          str = @reader.read_outer_xml
           @reader.next_sibling
+          return @product_klass.from_xml(str)
         end
       end
-      @queue.push nil
+      return nil
     end
 
+    # simple mapping of encoding constants to a string
+    #
     def encoding_const_to_name(const)
       case const
       when LibXML::XML::Encoding::UTF_8

data/spec/reader_spec.rb

@@ -10,6 +10,7 @@ context "ONIX::Reader" do
     data_path = File.join(File.dirname(__FILE__),"..","data")
     @file1    = File.join(data_path, "9780194351898.xml")
     @file2    = File.join(data_path, "two_products.xml")
+    @entity_file = File.join(data_path, "entities.xml")
   end
 
   specify "should initialize with a filename" do
@@ -17,14 +18,12 @@ context "ONIX::Reader" do
     reader.instance_variable_get("@reader").should be_a_kind_of(LibXML::XML::Reader)
   end
 
-=begin
   specify "should initialize with an IO object" do
     File.open(@file1,"rb") do |f|
       reader = ONIX::Reader.new(f)
       reader.instance_variable_get("@reader").should be_a_kind_of(LibXML::XML::Reader)
     end
   end
-=end
 
   specify "should provide access to various XML metadata from file" do
     reader = ONIX::Reader.new(@file1)
@@ -61,4 +60,20 @@ context "ONIX::Reader" do
     products[0].record_reference.should eql("365-9780194351898")
     products[1].record_reference.should eql("9780754672326")
   end
+
+  # libxml can handle the 3 standard entities fine (& < and ^gt;) but
+  # barfs when it encounters others. In theory other entityies are defined in the
+  # ONIX DTD, but I can't work out how to get libxml to recognise them
+  specify "should correctly parse a file that has an entity in it" do
+    reader = ONIX::Reader.new(@entity_file)
+    
+    products = []
+    reader.each do |product|
+      products << product
+    end
+
+    products.size.should eql(1)
+    products.first.record_reference.should eql("9780732287573")
+    products.first.titles.first.title_text.should eql("High Noon\342\200\223in Nimbin")
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: onix
 version: !ruby/object:Gem::Version 
-  version: 0.5.1
+  version: 0.6.0
 platform: ruby
 authors: 
 - James Healy
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-03-04 00:00:00 +11:00
+date: 2009-03-18 00:00:00 +11:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -20,7 +20,7 @@ dependencies:
     requirements: 
     - - "="
       - !ruby/object:Gem::Version 
-        version: 2.5.1
+        version: 2.5.2
     version: 
 - !ruby/object:Gem::Dependency 
   name: libxml-ruby