Fix parsing of invalid PDFs containing non-existing indirect stream l…

…ength values

CHANGELOG.md

@@ -20,6 +20,8 @@
 * [HexaPDF::DigitalSignature::Signing::DefaultHandler] to update the document's
   version to 2.0 when using PAdES
 * Parsing of invalid `)` character in PDF objects and content streams
+* Handling of files that contain stream length values that are indirect objects
+  that do not exist
 
 
 ## 0.47.0 - 2024-09-07

lib/hexapdf/parser.rb

@@ -184,7 +184,7 @@ def parse_indirect_object(offset = nil)
         length = if object[:Length].kind_of?(Integer)
                    object[:Length]
                  elsif object[:Length].kind_of?(Reference)
-                   @document.deref(object[:Length]).value
+                   @document.deref(object[:Length])&.value || 0
                  else
                    0
                  end

test/hexapdf/test_parser.rb

@@ -152,6 +152,13 @@ def create_parser(str)
       assert_equal('12', collector(stream.fiber))
     end
 
+    it "recovers from a non-existing indirect reference to a stream length value" do
+      create_parser("1 0 obj<</Length 2 0 R>> stream\n12(ab\nendstream endobj")
+      obj, _, _, stream = @parser.parse_indirect_object
+      assert_equal(5, obj[:Length])
+      assert_equal('12(ab', collector(stream.fiber))
+    end
+
     it "works even if the keyword endobj is missing or mangled" do
       create_parser("1 0 obj<</Length 4>>5")
       object, * = @parser.parse_indirect_object