Account unused object ids

lib/rubrik/document.rb

@@ -17,19 +17,25 @@ class Document
     sig {returns(PDF::Reader::ObjectHash)}
     attr_accessor :objects
 
+    sig {returns(Integer)}
+    attr_accessor :first_free_object_id
+
     sig {returns(T::Array[{id: PDF::Reader::Reference, value: T.untyped}])}
     attr_accessor :modified_objects
 
     sig {returns(Integer)}
     attr_accessor :last_object_id
 
-    private :io=, :objects=, :modified_objects=, :last_object_id=
+    private :io=, :objects=, :first_free_object_id=, :modified_objects=, :last_object_id=
 
     sig {params(input: T.any(File, Tempfile, StringIO)).void}
     def initialize(input)
       self.io = input
       self.objects = PDF::Reader::ObjectHash.new(input)
-      self.last_object_id = objects.size
+
+      self.last_object_id = objects.trailer[:Size] - 1
+      self.first_free_object_id = find_first_free_object_id
+
       self.modified_objects = []
 
       fetch_or_create_interactive_form!
@@ -87,6 +93,16 @@ def interactive_form
       T.must(modified_objects.first).fetch(:value)
     end
 
+    sig{returns(Integer)}
+    def find_first_free_object_id
+      return 0 if last_object_id == objects.size
+
+      xref = objects.send(:xref).instance_variable_get(:@xref)
+      missing_ids = (1..last_object_id).to_a - xref.keys
+
+      T.must(missing_ids.min)
+    end
+
     sig {void}
     def fetch_or_create_interactive_form!
       root_ref = objects.trailer[:Root]

lib/rubrik/document/increment.rb

@@ -46,7 +46,7 @@ def call(document, io:)
           io << "#{starting_id} #{length}\n"
 
           if starting_id.zero?
-            io << "0000000000 65535 f \n"
+            io << "#{format("%010d", document.first_free_object_id)} 65535 f \n"
             subsection.shift
           end
 

test/rubrik/document_test.rb

@@ -15,7 +15,7 @@ def test_initialize_document_without_interactive_form
       # Assert
       assert_equal(input, document.send(:io))
       # FACT: the interactive form was created
-      assert_equal(5, document.last_object_id)
+      assert_equal(6, document.last_object_id)
       assert_kind_of(PDF::Reader::ObjectHash, document.objects)
 
       acro_form = document.modified_objects.find { |obj| obj.dig(:value, :Type) == :Catalog }