specs/data/content_stream_with_length_as_ref_and_windows_breaks.pdf @@ -72,7 +72,7 @@ class PDF::Reader # If the object is a stream, that is returned as well def object (ref, save_pos = true) return ref unless ref.kind_of?(Reference) - pos = @buffer.pos if save_pos + pos = @buffer.pos - @buffer.raw.to_s.size if save_pos obj = Parser.new(@buffer.seek(offset_for(ref)), self).object(ref.id, ref.gen) @buffer.seek(pos) if save_pos return obj lib/pdf/reader/xref.rb @@ -135,4 +135,16 @@ context "PDF::Reader" do receiver.content.size.should eql(1) receiver.content[0].should eql("HelloWorld") end + + # PDF::Reader::XRef#object was saving an incorrect position when seeking. We + # were saving the current pos of the underlying IO stream, then seeking back + # to it. This was fine, except when there was still content in the buffer. + specify "should correctly process a PDF with a stream that has its length specified as an indirect reference and uses windows line breaks" do + receiver = PageTextReceiver.new + PDF::Reader.file(File.dirname(__FILE__) + "/data/content_stream_with_length_as_ref_and_windows_breaks.pdf", receiver) + + # confirm the text appears on the correct pages + receiver.content.size.should eql(1) + receiver.content[0].should eql("HelloWorld") + end end specs/meta_spec.rb ebd59ab016a51b1d057251f9624ad9cbcfa9cf03 James Healy jimmy@deefa.com http://github.com/yob/pdf-reader/commit/090638048a38896b2b501421837ebbb3be9d9b03 090638048a38896b2b501421837ebbb3be9d9b03 2009-04-24T01:59:31-07:00 2009-04-24T01:59:31-07:00 Fix a bug in PDF::Reader::XRef#object() - PDF::Reader::XRef#object was saving an incorrect position when seeking. We were saving the current pos of the underlying IO stream, then seeking back to it. This was fine, except when there was still content in the buffer. 967b8014d995fe3b6c41d6206614e30a0c491639 James Healy jimmy@deefa.com