Parse *ALL* content instructions when they're split over multiple con…

…tent streams

* thanks to Jack Rusher for reporting

CHANGELOG

@@ -1,6 +1,8 @@
 v0.8.6 (XXX)
 - initial support for pdf 1.5+ files that use object and xref streams
 - support streams that use a flate filter with the predictor option
+- ensure all content instructions are parsed when split over multiple stream
+  - thanks to Jack Rusher for reporting
 
 v0.8.5 (11th April 2010)
 - fix a regression introduced in 0.8.4.

lib/pdf/reader/content.rb

@@ -311,10 +311,8 @@ def walk_pages (page)
         fonts = font_hash_from_resources(current_resources)
 
         if page.has_key?(:Contents) and page[:Contents]
-          contents.each do |content|
-            obj = @xref.object(content)
-            content_stream(obj, fonts)
-          end 
+          direct_contents = contents.map { |content| @xref.object(content) }
+          content_stream(direct_contents, fonts)
         end
 
         resources.pop if res
@@ -353,7 +351,10 @@ def current_resources
     # Reads a PDF content stream and calls all the appropriate callback methods for the operators
     # it contains
     def content_stream (instructions, fonts = {})
-      instructions = instructions.unfiltered_data if instructions.kind_of?(PDF::Reader::Stream)
+      instructions = [instructions] unless instructions.kind_of?(Array)
+      instructions = instructions.map { |ins|
+        ins.is_a?(PDF::Reader::Stream) ? ins.unfiltered_data : ins.to_s
+      }.join
       buffer       = Buffer.new(StringIO.new(instructions))
       parser       = Parser.new(buffer, @xref)
       current_font = nil

specs/content_spec.rb

@@ -8,7 +8,7 @@ class PDF::Reader::XRef
   attr_accessor :xref
 end
 
-context "The PDF::Reader::Content class" do
+context PDF::Reader::Content do
 
   specify "should send the correct callbacks when processing instructions containing a single text block" do
 
@@ -71,6 +71,21 @@ class PDF::Reader::XRef
     content.content_stream(obj)
   end
 
+  # test for a bug reported by Jack Rusher where params at the end of a stream would be
+  # silently dropped if their matching operator was in the next contream stream in a series
+  specify "should send the correct callbacks when processing a PDF with content over multiple streams" do
+
+    receiver = PDF::Reader::RegisterReceiver.new
+
+    filename = File.dirname(__FILE__) + "/data/split_params_and_operator.pdf"
+    PDF::Reader.file(filename, receiver)
+
+    text_callbacks = receiver.all(:show_text_with_positioning)
+    text_callbacks.size.should eql(2)
+    text_callbacks[0][:args].should eql([["My name is"]])
+    text_callbacks[1][:args].should eql([["James Healy"]])
+  end
+
   specify "should send the correct metadata callbacks when processing an PrinceXML PDF" do
 
     receiver = PDF::Reader::RegisterReceiver.new

specs/meta_spec.rb

@@ -206,7 +206,17 @@ def show_text_with_positioning(*params)
     receiver = PageTextReceiver.new
     PDF::Reader.file(File.dirname(__FILE__) + "/data/indirect_xobject.pdf", receiver)
 
-    # confirm there was a single page of tet
+    # confirm there was a single page of text
     receiver.content.size.should eql(1)
   end
+
+  specify "should correctly process a PDF that uses multiple content streams for a single page" do
+    receiver = PageTextReceiver.new
+    PDF::Reader.file(File.dirname(__FILE__) + "/data/split_params_and_operator.pdf", receiver)
+
+    # confirm there was a single page of text
+    receiver.content.size.should eql(1)
+    receiver.content[0].include?("My name is").should be_true
+    receiver.content[0].include?("James Healy").should be_true
+  end
 end