hexapdf 0.12.3 → 0.13.0

data/lib/hexapdf/version.rb

@@ -37,6 +37,6 @@
 module HexaPDF
 
   # The version of HexaPDF.
-  VERSION = '0.12.3'
+  VERSION = '0.13.0'
 
 end

data/test/hexapdf/common_tokenizer_tests.rb

@@ -122,7 +122,7 @@ module CommonTokenizerTests
   end
 
   it "next_token: should not fail when reading super long numbers" do
-    create_tokenizer("1" + "0" * 10_000)
+    create_tokenizer("1" << "0" * 10_000)
     assert_equal(10**10_000, @tokenizer.next_token)
   end
 
@@ -162,7 +162,7 @@ module CommonTokenizerTests
   end
 
   it "returns the correct position on operations" do
-    create_tokenizer("hallo du" + " " * 50000 + "hallo du")
+    create_tokenizer("hallo du" << " " * 50000 << "hallo du")
     @tokenizer.next_token
     assert_equal(5, @tokenizer.pos)
 

data/test/hexapdf/content/graphic_object/test_arc.rb

@@ -68,14 +68,14 @@ describe HexaPDF::Content::GraphicObject::Arc do
       arc.max_curves = 4
       curves = arc.curves
       assert_equal(2, curves.size)
-      assert_curve_values([0, 1, p1: [1, 0.548584], p2: [0.548584, 1]], curves[0])
-      assert_curve_values([-1, 0, p1: [-0.548584, 1], p2: [-1, 0.548584]], curves[1])
+      assert_curve_values([0, 1, {p1: [1, 0.548584], p2: [0.548584, 1]}], curves[0])
+      assert_curve_values([-1, 0, {p1: [-0.548584, 1], p2: [-1, 0.548584]}], curves[1])
 
       arc.configure(clockwise: true)
       curves = arc.curves
       assert_equal(2, curves.size)
-      assert_curve_values([0, -1, p1: [1, -0.548584], p2: [0.548584, -1]], curves[0])
-      assert_curve_values([-1, 0, p1: [-0.548584, -1], p2: [-1, -0.548584]], curves[1])
+      assert_curve_values([0, -1, {p1: [1, -0.548584], p2: [0.548584, -1]}], curves[0])
+      assert_curve_values([-1, 0, {p1: [-0.548584, -1], p2: [-1, -0.548584]}], curves[1])
     end
   end
 

data/test/hexapdf/content/test_canvas.rb

@@ -531,7 +531,7 @@ describe HexaPDF::Content::Canvas do
     end
 
     it "invokes the polygon method when radius != 0" do
-      args = [0, 0, 10, 0, 10, 10, 0, 10, radius: 5]
+      args = [0, 0, 10, 0, 10, 10, 0, 10, {radius: 5}]
       assert_method_invoked(@canvas, :polygon, args) do
         @canvas.rectangle(0, 0, 10, 10, radius: 5)
       end
@@ -631,7 +631,7 @@ describe HexaPDF::Content::Canvas do
 
   describe "circle" do
     it "uses arc for the hard work" do
-      assert_method_invoked(@canvas, :arc, [5, 6, a: 7]) do
+      assert_method_invoked(@canvas, :arc, [5, 6, {a: 7}]) do
         @canvas.graphics_object = :path
         @canvas.circle(5, 6, 7)
       end
@@ -651,7 +651,7 @@ describe HexaPDF::Content::Canvas do
 
   describe "ellipse" do
     it "uses arc for the hard work" do
-      assert_method_invoked(@canvas, :ellipse, [5, 6, a: 7, b: 5, inclination: 10]) do
+      assert_method_invoked(@canvas, :ellipse, [5, 6, {a: 7, b: 5, inclination: 10}]) do
         @canvas.ellipse(5, 6, a: 7, b: 5, inclination: 10)
       end
     end

data/test/hexapdf/content/test_color_space.rb

@@ -155,7 +155,7 @@ describe HexaPDF::Content::ColorSpace::DeviceGray do
 
   before do
     @color_space = HexaPDF::Content::ColorSpace::DeviceGray.new
-    @color_space_family = @color_space_definition =  :DeviceGray
+    @color_space_family = @color_space_definition = :DeviceGray
     @color = @color_space.default_color
     @other_color = @color_space.color(128)
     @colors = [128]

data/test/hexapdf/encryption/test_aes.rb

@@ -101,13 +101,13 @@ describe HexaPDF::Encryption::AES do
       result = TestHelper.collector(@algorithm_class.decryption_fiber('some' * 4, f))
       assert_equal('a' * 16, result)
 
-      f = Fiber.new { 'a' * 31 + "\x00" }
+      f = Fiber.new { 'a' * 31 << "\x00" }
       result = TestHelper.collector(@algorithm_class.decryption_fiber('some' * 4, f))
-      assert_equal('a' * 15 + "\x00", result)
+      assert_equal('a' * 15 << "\x00", result)
 
-      f = Fiber.new { 'a' * 29 + "\x00\x01\x03" }
+      f = Fiber.new { 'a' * 29 << "\x00\x01\x03" }
       result = TestHelper.collector(@algorithm_class.decryption_fiber('some' * 4, f))
-      assert_equal('a' * 13 + "\x00\x01\x03", result)
+      assert_equal('a' * 13 << "\x00\x01\x03", result)
     end
 
     it "fails on decryption if not enough bytes are provided" do

data/test/hexapdf/encryption/test_standard_security_handler.rb

@@ -53,24 +53,24 @@ describe HexaPDF::Encryption::StandardEncryptionDictionary do
 end
 
 describe HexaPDF::Encryption::StandardSecurityHandler do
-  TEST_FILES = Dir[File.join(TEST_DATA_DIR, 'standard-security-handler', '*.pdf')].sort
-  USER_PASSWORD = 'uhexapdf'
-  OWNER_PASSWORD = 'ohexapdf'
+  test_files = Dir[File.join(TEST_DATA_DIR, 'standard-security-handler', '*.pdf')].sort
+  user_password = 'uhexapdf'
+  owner_password = 'ohexapdf'
 
-  MINIMAL_DOC = HexaPDF::Document.new(io: StringIO.new(MINIMAL_PDF))
+  minimal_doc = HexaPDF::Document.new(io: StringIO.new(MINIMAL_PDF))
 
-  TEST_FILES.each do |file|
+  test_files.each do |file|
     basename = File.basename(file)
     it "can decrypt, encrypt and decrypt the encrypted file #{basename} with the user password" do
       begin
         doc = HexaPDF::Document.new(io: StringIO.new(File.binread(file)),
-                                    decryption_opts: {password: USER_PASSWORD})
-        assert_equal(MINIMAL_DOC.trailer[:Info][:ModDate], doc.trailer[:Info][:ModDate])
+                                    decryption_opts: {password: user_password})
+        assert_equal(minimal_doc.trailer[:Info][:ModDate], doc.trailer[:Info][:ModDate])
 
         out = StringIO.new(''.b)
         HexaPDF::Writer.new(doc, out).write
-        doc = HexaPDF::Document.new(io: out, decryption_opts: {password: USER_PASSWORD})
-        assert_equal(MINIMAL_DOC.trailer[:Info][:ModDate], doc.trailer[:Info][:ModDate])
+        doc = HexaPDF::Document.new(io: out, decryption_opts: {password: user_password})
+        assert_equal(minimal_doc.trailer[:Info][:ModDate], doc.trailer[:Info][:ModDate])
       rescue HexaPDF::EncryptionError => e
         flunk("Error processing #{basename}: #{e}")
       end
@@ -80,8 +80,8 @@ describe HexaPDF::Encryption::StandardSecurityHandler do
       it "can decrypt the encrypted file #{basename} with the owner password" do
         begin
           doc = HexaPDF::Document.new(io: StringIO.new(File.binread(file)),
-                                      decryption_opts: {password: OWNER_PASSWORD})
-          assert_equal(MINIMAL_DOC.trailer[:Info][:ModDate], doc.trailer[:Info][:ModDate])
+                                      decryption_opts: {password: owner_password})
+          assert_equal(minimal_doc.trailer[:Info][:ModDate], doc.trailer[:Info][:ModDate])
         rescue HexaPDF::EncryptionError => e
           flunk("Error processing #{basename}: #{e}")
         end

data/test/hexapdf/filter/test_ascii85_decode.rb

@@ -33,7 +33,7 @@ describe HexaPDF::Filter::ASCII85Decode do
     end
 
     it "ignores data after the EOD marker" do
-      assert_equal(@decoded, collector(@obj.decoder(feeder(@encoded.dup + "~>abcdefg"))))
+      assert_equal(@decoded, collector(@obj.decoder(feeder(@encoded << "~>abcdefg"))))
     end
 
     it "fails if the input contains invalid characters" do

data/test/hexapdf/filter/test_ascii_hex_decode.rb

@@ -24,7 +24,7 @@ describe HexaPDF::Filter::ASCIIHexDecode do
     end
 
     it "ignores data after the EOD marker" do
-      assert_equal(@decoded, collector(@obj.decoder(feeder(@encoded + '4e6f7gzz'))))
+      assert_equal(@decoded, collector(@obj.decoder(feeder(@encoded << '4e6f7gzz'))))
     end
 
     it "assumes the missing char is '0' if the input length is odd" do

data/test/hexapdf/layout/test_text_layouter.rb

@@ -674,10 +674,9 @@ describe HexaPDF::Layout::TextLayouter do
       pos = [0, 0]
       result.select! {|name, _| name == :set_text_matrix || name == :move_text_next_line }.
         map! do |name, ops|
-        if name == :set_text_matrix
-          pos = ops[-2, 2]
-        elsif name == :move_text_next_line
-          pos[1] -= leading
+        case name
+        when :set_text_matrix then pos = ops[-2, 2]
+        when :move_text_next_line then pos[1] -= leading
         end
         pos.dup
       end

data/test/hexapdf/test_configuration.rb

@@ -66,8 +66,8 @@ describe HexaPDF::Configuration do
       assert_equal(HexaPDF, @config.constantize('test', 1))
     end
 
-    def assert_constantize_error # :nodoc:
-      exp = assert_raises(HexaPDF::Error) { yield }
+    def assert_constantize_error(&block) # :nodoc:
+      exp = assert_raises(HexaPDF::Error, &block)
       assert_match(/Error getting constant for configuration option/, exp.message)
     end
 

data/test/hexapdf/test_dictionary.rb

@@ -14,7 +14,9 @@ describe HexaPDF::Dictionary do
   end
 
   def add(obj)
-    HexaPDF::Object.new(obj, oid: 1)
+    klass = HexaPDF::Object
+    klass = HexaPDF::Dictionary if obj.kind_of?(HexaPDF::Dictionary) || obj.kind_of?(Hash)
+    klass.new(obj, oid: 1)
   end
 
   def delete(_obj)

data/test/hexapdf/test_dictionary_fields.rb

@@ -222,7 +222,7 @@ describe HexaPDF::DictionaryFields do
 
     it "allows conversion to a Rectangle from an Array" do
       doc = Minitest::Mock.new
-      doc.expect(:wrap, :data, [[0, 1, 2, 3], type: HexaPDF::Rectangle])
+      doc.expect(:wrap, :data, [[0, 1, 2, 3], {type: HexaPDF::Rectangle}])
       @field.convert([0, 1, 2, 3], doc)
       doc.verify
     end
@@ -230,7 +230,7 @@ describe HexaPDF::DictionaryFields do
     it "allows conversion to a Rectangle from a HexaPDF::PDFArray" do
       data = HexaPDF::PDFArray.new([0, 1, 2, 3])
       doc = Minitest::Mock.new
-      doc.expect(:wrap, :data, [data, type: HexaPDF::Rectangle])
+      doc.expect(:wrap, :data, [data, {type: HexaPDF::Rectangle}])
       @field.convert(data, doc)
       doc.verify
     end

data/test/hexapdf/test_document.rb

@@ -441,21 +441,21 @@ describe HexaPDF::Document do
 
   describe "validate" do
     before do
-      @doc.trailer.validate # to create a valid document
+      @doc.validate # to create a valid document
     end
 
     it "validates indirect objects" do
-      obj = @doc.add({Type: :Catalog})
+      obj = @doc.add({Type: :Page, MediaBox: [1, 1, 1, 1], Parent: @doc.pages.root})
       refute(@doc.validate(auto_correct: false))
 
       called = false
-      assert(@doc.validate {|o| assert_same(obj, o); called = true })
+      assert(@doc.validate {|_, _, o| assert_same(obj, o); called = true })
       assert(called)
     end
 
     it "validates the trailer object" do
       @doc.trailer[:ID] = :Symbol
-      refute(@doc.validate {|obj| assert_same(@doc.trailer, obj) })
+      refute(@doc.validate {|_, _, obj| assert_same(@doc.trailer, obj) })
     end
 
     it "validates only loaded objects" do

data/test/hexapdf/test_object.rb

@@ -6,15 +6,6 @@ require 'hexapdf/reference'
 
 describe HexaPDF::Object do
   describe "class.deep_copy" do
-    it "handles not-duplicatable classes" do
-      assert_equal(5, HexaPDF::Object.deep_copy(5))
-      assert_equal(5.5, HexaPDF::Object.deep_copy(5.5))
-      assert_nil(HexaPDF::Object.deep_copy(nil))
-      assert_equal(true, HexaPDF::Object.deep_copy(true))
-      assert_equal(false, HexaPDF::Object.deep_copy(false))
-      assert_equal(:Name, HexaPDF::Object.deep_copy(:Name))
-    end
-
     it "handles general, duplicatable classes" do
       x = "test"
       assert_equal("test", HexaPDF::Object.deep_copy(x))
@@ -103,30 +94,57 @@ describe HexaPDF::Object do
   end
 
   describe "validate" do
-    it "invokes perform_validation correctly via #validate" do
-      obj = HexaPDF::Object.new(5)
-      invoked = {}
-      obj.define_singleton_method(:perform_validation) do |&block|
-        invoked[:method] = true
+    before do
+      @obj = HexaPDF::Object.new(5)
+    end
+
+    it "invokes perform_validation correctly" do
+      invoked = false
+      @obj.define_singleton_method(:perform_validation) { invoked = true }
+      assert(@obj.validate)
+      assert(invoked)
+    end
+
+    it "yields all arguments yieled by perform_validation" do
+      invoked = []
+      @obj.define_singleton_method(:perform_validation) do |&block|
+        block.call("error", true, :object)
+      end
+      assert(@obj.validate {|*a| invoked << a })
+      assert_equal([["error", true, :object]], invoked)
+    end
+
+    it "provides self as third argument if none is yielded by perform_validation" do
+      invoked = []
+      @obj.define_singleton_method(:perform_validation) do |&block|
         block.call("error", true)
       end
-      assert(obj.validate {|*a| invoked[:block] = a })
-      assert_equal([:method, :block], invoked.keys)
-      assert_equal(["error", true], invoked[:block])
+      assert(@obj.validate {|*a| invoked << a })
+      assert_equal([["error", true, @obj]], invoked)
+    end
 
-      refute(obj.validate(auto_correct: false))
+    it "yields all problems when auto_correct is true" do
+      invoked = []
+      @obj.define_singleton_method(:perform_validation) do |&block|
+        invoked << :before
+        block.call("error", false)
+        invoked << :after
+        block.call("error2", true)
+        invoked << :last
+      end
+      refute(@obj.validate)
+      assert_equal([:before, :after, :last], invoked)
     end
 
-    it "stops validating on an uncorrectable problem" do
-      obj = HexaPDF::Object.new(5)
-      invoked = {}
-      obj.define_singleton_method(:perform_validation) do |&block|
-        invoked[:before] = true
+    it "stops at the first uncorrectable problem if auto_correct is false" do
+      invoked = []
+      @obj.define_singleton_method(:perform_validation) do |&block|
+        invoked << :before
         block.call("error", false)
-        invoked[:after] = true
+        invoked << :after
       end
-      refute(obj.validate {|*a| invoked[:block] = a })
-      refute(invoked.key?(:after))
+      refute(@obj.validate(auto_correct: false))
+      assert_equal([:before], invoked)
     end
   end
 

data/test/hexapdf/test_parser.rb

@@ -8,6 +8,7 @@ require 'stringio'
 describe HexaPDF::Parser do
   before do
     @document = HexaPDF::Document.new
+    @document.config['parser.try_xref_reconstruction'] = false
     @document.add(@document.wrap(10, oid: 1, gen: 0))
 
     create_parser(<<~EOF)
@@ -132,6 +133,48 @@ describe HexaPDF::Parser do
       exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object(0) }
       assert_match(/stream.*followed by.*endstream/i, exp.message)
     end
+
+    describe "with strict parsing" do
+      before do
+        @document.config['parser.on_correctable_error'] = proc { true }
+      end
+
+      it "fails if an empty indirect object is found" do
+        create_parser("1 0 obj\nendobj")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
+        assert_match(/no indirect object value/i, exp.message)
+      end
+
+      it "fails if keyword stream is followed only by CR without LF" do
+        create_parser("1 0 obj<</Length 2>> stream\r12\nendstream endobj")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
+        assert_match(/not CR alone/, exp.message)
+      end
+
+      it "fails if the stream length value is invalid" do
+        create_parser("1 0 obj<</Length 4>> stream\n12endstream endobj")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
+        assert_match(/invalid stream length/i, exp.message)
+      end
+
+      it "fails if the keyword endobj is mangled" do
+        create_parser("1 0 obj\n<< >>\nendobjd\n")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
+        assert_match(/keyword endobj/, exp.message)
+      end
+
+      it "fails if the keyword endobj is missing" do
+        create_parser("1 0 obj\n<< >>")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
+        assert_match(/keyword endobj/, exp.message)
+      end
+
+      it "fails if there is data between 'endstream' and 'endobj'" do
+        create_parser("1 0 obj\n<< >>\nstream\nendstream\ntest\nendobj\n")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object(0) }
+        assert_match(/keyword endobj/, exp.message)
+      end
+    end
   end
 
   describe "load_object" do
@@ -205,7 +248,7 @@ describe HexaPDF::Parser do
     end
 
     it "ignores garbage at the end of the file" do
-      create_parser("startxref\n5\n%%EOF" + "\nhallo" * 150)
+      create_parser("startxref\n5\n%%EOF" << "\nhallo" * 150)
       assert_equal(5, @parser.startxref_offset)
     end
 
@@ -215,9 +258,9 @@ describe HexaPDF::Parser do
     end
 
     it "finds the startxref anywhere in file" do
-      create_parser("startxref\n5\n%%EOF" + "\nhallo" * 5000)
+      create_parser("startxref\n5\n%%EOF" << "\nhallo" * 5000)
       assert_equal(5, @parser.startxref_offset)
-      create_parser("startxref\n5\n%%EOF\n" + "h" * 1017)
+      create_parser("startxref\n5\n%%EOF\n" << "h" * 1017)
       assert_equal(5, @parser.startxref_offset)
     end
 
@@ -242,6 +285,13 @@ describe HexaPDF::Parser do
       exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.startxref_offset }
       assert_match(/missing startxref/, exp.message)
     end
+
+    it "fails on strict parsing if the startxref is not in the last part of the file" do
+      @document.config['parser.on_correctable_error'] = proc { true }
+      create_parser("startxref\n5\n%%EOF" << "\nhallo" * 5000)
+      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.startxref_offset }
+      assert_match(/end-of-file marker not found/, exp.message)
+    end
   end
 
   describe "file_header_version" do
@@ -262,7 +312,7 @@ describe HexaPDF::Parser do
     end
 
     it "ignores junk at the beginning of the file and correctly calculates offset" do
-      create_parser("junk" * 200 + "\n%PDF-1.4\n")
+      create_parser("junk" * 200 << "\n%PDF-1.4\n")
       assert_equal('1.4', @parser.file_header_version)
       assert_equal(801, @parser.instance_variable_get(:@header_offset))
     end
@@ -318,6 +368,12 @@ describe HexaPDF::Parser do
       assert_match(/invalid cross-reference subsection/i, exp.message)
     end
 
+    it "fails if a sub section entry is mangled" do
+      create_parser("xref\n0 2\n000a000000 00000 n\n0000000000 65535 n\ntrailer\n<<>>\n")
+      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
+      assert_match(/invalid cross-reference entry/i, exp.message)
+    end
+
     it "fails if there is no trailer" do
       create_parser("xref\n0 1\n0000000000 00000 n \n")
       exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
@@ -329,6 +385,30 @@ describe HexaPDF::Parser do
       exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
       assert_match(/dictionary/, exp.message)
     end
+
+    describe "with strict parsing" do
+      before do
+        @document.config['parser.on_correctable_error'] = proc { true }
+      end
+
+      it "fails if xref type=n with offset=0" do
+        create_parser("xref\n0 2\n0000000000 00000 n \n0000000000 00000 n \ntrailer\n<<>>\n")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
+        assert_match(/invalid.*cross-reference entry/i, exp.message)
+      end
+
+      it " fails xref type=n with gen>65535" do
+        create_parser("xref\n0 2\n0000000000 00000 n \n0000000000 65536 n \ntrailer\n<<>>\n")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
+        assert_match(/invalid.*cross-reference entry/i, exp.message)
+      end
+
+      it "fails if trailing second whitespace is missing" do
+        create_parser("xref\n0 1\n0000000000 00000 n\ntrailer\n<<>>\n")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
+        assert_match(/invalid.*cross-reference entry/i, exp.message)
+      end
+    end
   end
 
   describe "load_revision" do
@@ -348,75 +428,55 @@ describe HexaPDF::Parser do
       exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.load_revision(10) }
       assert_match(/not a cross-reference stream/, exp.message)
     end
-  end
 
-  describe "with strict parsing enabled" do
-    before do
+    it "fails on strict parsing if the cross-reference stream doesn't contain an entry for itself" do
       @document.config['parser.on_correctable_error'] = proc { true }
+      create_parser("2 0 obj\n<</Type/XRef/Length 3/W [1 1 1]/Size 1>>" \
+                    "stream\n\x01\x0A\x00\nendstream endobj")
+      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.load_revision(0) }
+      assert_match(/entry for itself/, exp.message)
     end
+  end
 
-    it "startxref_offset fails if the startxref is not in the last part of the file" do
-      create_parser("startxref\n5\n%%EOF" + "\nhallo" * 5000)
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.startxref_offset }
-      assert_match(/end-of-file marker not found/, exp.message)
-    end
-
-    it "parse_xref_section_and_trailer fails if xref type=n with offset=0" do
-      create_parser("xref\n0 2\n0000000000 00000 n \n0000000000 00000 n \ntrailer\n<<>>\n")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
-      assert_match(/invalid.*cross-reference entry/i, exp.message)
-    end
-
-    it "parse_xref_section_and_trailer fails xref type=n with gen>65535" do
-      create_parser("xref\n0 2\n0000000000 00000 n \n0000000000 65536 n \ntrailer\n<<>>\n")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
-      assert_match(/invalid.*cross-reference entry/i, exp.message)
+  describe "reconstruct_revision" do
+    before do
+      @document.config['parser.try_xref_reconstruction'] = true
+      @xref = HexaPDF::XRefSection.in_use_entry(1, 0, 100)
     end
 
-    it "parse_xref_section_and_trailer fails if trailing second whitespace is missing" do
-      create_parser("xref\n0 1\n0000000000 00000 n\ntrailer\n<<>>\n")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_xref_section_and_trailer(0) }
-      assert_match(/invalid.*cross-reference subsection entry/i, exp.message)
+    it "serially parses the contents" do
+      create_parser("1 0 obj\n5\nendobj\n1 0 obj\n6\nendobj\ntrailer\n<</Size 1>>")
+      assert_equal(6, @parser.load_object(@xref).value)
     end
 
-    it "parse_indirect_object fails if an empty indirect object is found" do
-      create_parser("1 0 obj\nendobj")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
-      assert_match(/no indirect object value/i, exp.message)
+    it "ignores parts where the starting line is split across lines" do
+      create_parser("1 0 obj\n5\nendobj\n1 0\nobj\n6\nendobj\ntrailer\n<</Size 1>>")
+      assert_equal(5, @parser.load_object(@xref).value)
     end
 
-    it "parse_indirect_object fails if keyword stream is followed only by CR without LF" do
-      create_parser("1 0 obj<</Length 2>> stream\r12\nendstream endobj")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
-      assert_match(/not CR alone/, exp.message)
+    it "ignores invalid objects" do
+      create_parser("1 x obj\n5\nendobj\n1 0 xobj\n6\nendobj\n1 0 obj 4\nendobj\ntrailer\n<</Size 1>>")
+      assert_equal(4, @parser.load_object(@xref).value)
     end
 
-    it "parse_indirect_object fails if the stream length value is invalid" do
-      create_parser("1 0 obj<</Length 4>> stream\n12endstream endobj")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
-      assert_match(/invalid stream length/i, exp.message)
+    it "ignores invalid lines" do
+      create_parser("1 0 obj\n5\nendobj\nhello there\n1 0 obj\n6\nendobj\ntrailer\n<</Size 1>>")
+      assert_equal(6, @parser.load_object(@xref).value)
     end
 
-    it "parse_indirect_object fails if the keyword endobj is missing or mangled" do
-      create_parser("1 0 obj\n<< >>\nendobjd\n")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
-      assert_match(/keyword endobj/, exp.message)
-      create_parser("1 0 obj\n<< >>")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
-      assert_match(/keyword endobj/, exp.message)
+    it "uses the last trailer" do
+      create_parser("trailer <</Size 1>>\ntrailer <</Size 2/Prev 342>>")
+      assert_equal({Size: 2}, @parser.reconstructed_revision.trailer.value)
     end
 
-    it "parse_indirect_object fails if there is data between 'endstream' and 'endobj'" do
-      create_parser("1 0 obj\n<< >>\nstream\nendstream\ntest\nendobj\n")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object(0) }
-      assert_match(/keyword endobj/, exp.message)
+    it "uses the first trailer in case of a linearized file" do
+      create_parser("trailer <</Size 1/Prev 342>>\ntrailer <</Size 2>>")
+      assert_equal({Size: 1}, @parser.reconstructed_revision.trailer.value)
     end
 
-    it "load_revision fails if the cross-reference stream doesn't contain an entry for itself" do
-      create_parser("2 0 obj\n<</Type/XRef/Length 3/W [1 1 1]/Size 1>>" \
-                    "stream\n\x01\x0A\x00\nendstream endobj")
-      exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.load_revision(0) }
-      assert_match(/entry for itself/, exp.message)
+    it "fails if no valid trailer is found" do
+      create_parser("1 0 obj\n5\nendobj")
+      assert_raises(HexaPDF::MalformedPDFError) { @parser.load_object(@xref) }
     end
   end
 end