hexapdf 0.15.5 → 0.15.9

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6aa5d94e3b69f1d7b2b369d248664c2c4d96d13306c1fbb1cf0960412f129faf
-  data.tar.gz: f6e8ae5f61de78c41fecace64d0ad5d42f8c46032e6ac6550c01b2fb43109fe8
+  metadata.gz: 24d17dfd6c8dc9e3f7014e1ea769dede6f8cea81529bb201a8447f21873d3b25
+  data.tar.gz: 7165a2e11983731ba2597d2e4a824415abc96936d2b58f3fb737a4fed94dcf16
 SHA512:
-  metadata.gz: 49f3b3aee754308956df811f564d127269e8e1dbfc613bed39e42bd77303855b80fe11c2ade362cef174838b58188470b5604e2d2c930dcf7637e7600410ac15
-  data.tar.gz: 3a4d5555ae6f012176aebe0a29c36500ad20f385a342c27819f5db3a0e8c5e270c58526c8434b8c8d98c59b963721a51344e93bfc6d51f3eab4fc76cceeb830e
+  metadata.gz: 244332a4f024c90cf6344b462ed422a5f73b32d3a4d04d0dcdadc6e7ede2cd0724f7a3329fa3aad68c99e47f49a9dff66806d9b8b152b9551eaecd7365c807d4
+  data.tar.gz: 06c4b9fd5ecd8f045a37e85ce4a6539f52b48bb16fb685290bcc2a0210b5f2e66bcfa9b7efb1115c45dd1618de9421c479d7a66d84556d1b9c37b19b1f8b6075

data/CHANGELOG.md

@@ -1,3 +1,34 @@
+## 0.15.9 - 2021-09-04
+
+### Fixed
+
+* Handling of files that contain stream length values that are indirect objects
+  not referring to a number
+
+
+## 0.15.8 - 2021-08-16
+
+### Fixed
+
+* Regression when using `-v` with the hexapdf command line tool
+
+
+## 0.15.7 - 2021-07-17
+
+### Fixed
+
+* Infinite loop while parsing PDF array due to missing closing bracket
+* Handling of invalid files with missing or corrupted trailer dictionary
+
+
+## 0.15.6 - 2021-07-16
+
+### Fixed
+
+* Handling of indirect objects with invalid values which are now treated as null
+  objects
+
+
 ## 0.15.5 - 2021-07-06
 
 ### Changed

data/lib/hexapdf/cli/command.rb

@@ -119,7 +119,7 @@ module HexaPDF
       # Writes the document to the given file or does nothing if +out_file+ is +nil+.
       def write_document(doc, out_file, incremental: false)
         if out_file
-          doc.validate(auto_correct: true) do |object, msg, correctable|
+          doc.validate(auto_correct: true) do |msg, correctable, object|
             if command_parser.strict && !correctable
               raise "Validation error for object (#{object.oid},#{object.gen}): #{msg}"
             elsif command_parser.verbosity_info?

data/lib/hexapdf/parser.rb

@@ -125,11 +125,14 @@ module HexaPDF
         begin
           object = @tokenizer.next_object
         rescue MalformedPDFError
-          # Handle often found invalid indirect object with missing whitespace after number
-          maybe_raise("Invalid object value after 'obj'", pos: @tokenizer.pos,
-                      force: !(tok.kind_of?(Tokenizer::Token) && tok =~ /\A\d+endobj\z/))
-          object = tok.to_i
-          @tokenizer.pos -= 6
+          if tok.kind_of?(Tokenizer::Token) && tok =~ /\A\d+endobj\z/
+            # Handle often found invalid indirect object with missing whitespace after number
+            maybe_raise("Missing whitespace after number'", pos: @tokenizer.pos)
+            object = tok.to_i
+            @tokenizer.pos -= 6
+          else
+            maybe_raise("Invalid value after '#{oid} #{gen} obj', treating as null", pos: @tokenizer.pos)
+          end
         end
       end
 
@@ -162,7 +165,7 @@ module HexaPDF
                  else
                    0
                  end
-        @tokenizer.pos = pos + length
+        @tokenizer.pos = pos + length rescue pos
 
         tok = @tokenizer.next_token
         unless tok.kind_of?(Tokenizer::Token) && tok == 'endstream'
@@ -444,6 +447,15 @@ module HexaPDF
 
       if !trailer || trailer.empty?
         _, trailer = load_revision(startxref_offset) rescue nil
+        unless trailer
+          xref.each do |_oid, _gen, xref_entry|
+            obj, * = parse_indirect_object(xref_entry.pos) rescue nil
+            if obj.kind_of?(Hash) && obj[:Type] == :Catalog
+              trailer = {Root: HexaPDF::Reference.new(xref_entry.oid, xref_entry.gen)}
+              break
+            end
+          end
+        end
         unless trailer
           @in_reconstruct_revision = false
           raise_malformed("Could not reconstruct malformed PDF because trailer was not found", pos: 0)

data/lib/hexapdf/tokenizer.rb

@@ -55,6 +55,9 @@ module HexaPDF
 
     # This object is returned when there are no more tokens to read.
     NO_MORE_TOKENS = ::Object.new
+    def NO_MORE_TOKENS.to_s
+      "EOS - no more tokens"
+    end
 
     # Characters defined as whitespace.
     #
@@ -384,7 +387,11 @@ module HexaPDF
       result = []
       while true
         obj = next_object(allow_end_array_token: true)
-        break if obj.equal?(TOKEN_ARRAY_END)
+        if obj.equal?(TOKEN_ARRAY_END)
+          break
+        elsif obj.equal?(NO_MORE_TOKENS)
+          raise HexaPDF::MalformedPDFError.new("Unclosed array found", pos: pos)
+        end
         result << obj
       end
       result
@@ -403,7 +410,8 @@ module HexaPDF
         key = next_token
         break if key.equal?(TOKEN_DICT_END)
         unless key.kind_of?(Symbol)
-          raise HexaPDF::MalformedPDFError.new("Dictionary keys must be PDF name objects", pos: pos)
+          raise HexaPDF::MalformedPDFError.new("Dictionary keys must be PDF name objects, " \
+                                               "found '#{key}'", pos: pos)
         end
 
         val = next_object

data/lib/hexapdf/version.rb

@@ -37,6 +37,6 @@
 module HexaPDF
 
   # The version of HexaPDF.
-  VERSION = '0.15.5'
+  VERSION = '0.15.9'
 
 end

data/test/hexapdf/common_tokenizer_tests.rb

@@ -161,6 +161,21 @@ module CommonTokenizerTests
     assert_raises(HexaPDF::MalformedPDFError) { @tokenizer.next_object }
   end
 
+  it "next_object: fails for an array without closing bracket, encountering EOS" do
+    create_tokenizer("[1 2")
+    exception = assert_raises(HexaPDF::MalformedPDFError) { @tokenizer.next_object }
+    assert_match(/Unclosed array found/, exception.message)
+  end
+
+  it "next_object: fails for a dictionary without closing bracket, encountering EOS" do
+    create_tokenizer("<</Name 5")
+    exception = assert_raises(HexaPDF::MalformedPDFError) { @tokenizer.next_object }
+    assert_match(/must be PDF name objects.*EOS/, exception.message)
+    create_tokenizer("<</Name 5 /Other")
+    exception = assert_raises(HexaPDF::MalformedPDFError) { @tokenizer.next_object }
+    assert_match(/must be PDF name objects.*EOS/, exception.message)
+  end
+
   it "returns the correct position on operations" do
     create_tokenizer("hallo du" << " " * 50000 << "hallo du")
     @tokenizer.next_token

data/test/hexapdf/test_parser.rb

@@ -107,13 +107,27 @@ describe HexaPDF::Parser do
       assert_equal(749, object)
     end
 
-    it "recovers from an invalid stream length value" do
+    it "treats indirect objects with invalid values as null objects" do
+      create_parser("1 0 obj <</test ( /other (end)>> endobj")
+      object, * =  @parser.parse_indirect_object
+      assert_nil(object)
+    end
+
+    it "recovers from a stream length value that doesn't reflect the correct length" do
       create_parser("1 0 obj<</Length 4>> stream\n12endstream endobj")
       obj, _, _, stream = @parser.parse_indirect_object
       assert_equal(2, obj[:Length])
       assert_equal('12', TestHelper.collector(stream.fiber))
     end
 
+    it "recovers from an invalid stream length value" do
+      create_parser("1 0 obj<</Length 2 0 R>> stream\n12endstream endobj")
+      @document.add([5], oid: 2)
+      obj, _, _, stream = @parser.parse_indirect_object
+      assert_equal(2, obj[:Length])
+      assert_equal('12', TestHelper.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
@@ -185,7 +199,13 @@ describe HexaPDF::Parser do
       it "fails for numbers followed by endobj without space" do
         create_parser("1 0 obj 749endobj")
         exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
-        assert_match(/Invalid object value after 'obj'/, exp.message)
+        assert_match(/Missing whitespace after number/, exp.message)
+      end
+
+      it "fails for invalid values" do
+        create_parser("1 0 obj <</test ( /other (end)>> endobj")
+        exp = assert_raises(HexaPDF::MalformedPDFError) { @parser.parse_indirect_object }
+        assert_match(/Invalid value after '1 0 obj'/, exp.message)
       end
 
       it "fails if the stream length value is invalid" do
@@ -607,7 +627,12 @@ describe HexaPDF::Parser do
       assert_equal({Size: 1}, @parser.reconstructed_revision.trailer.value)
     end
 
-    it "fails if no trailer is found and the trailer specified at the startxref position is not valid" do
+    it "constructs a trailer with a /Root entry if no valid trailer was found" do
+      create_parser("1 0 obj\n<</Type /Catalog/Pages 2 0 R>>\nendobj\nxref trailer <</Size 1/Prev 5\n%%EOF")
+      assert_equal({Root: HexaPDF::Reference.new(1, 0)}, @parser.reconstructed_revision.trailer.value)
+    end
+
+    it "fails if no valid trailer is found and couldn't be constructed" do
       create_parser("1 0 obj\n5\nendobj\nquack trailer <</Size 1>>\nstartxref\n22\n%%EOF")
       assert_raises(HexaPDF::MalformedPDFError) { @parser.reconstructed_revision.trailer }
     end

data/test/hexapdf/test_writer.rb

@@ -40,7 +40,7 @@ describe HexaPDF::Writer do
       219
       %%EOF
       3 0 obj
-      <</Producer(HexaPDF version 0.15.5)>>
+      <</Producer(HexaPDF version 0.15.9)>>
       endobj
       xref
       3 1
@@ -72,7 +72,7 @@ describe HexaPDF::Writer do
       141
       %%EOF
       6 0 obj
-      <</Producer(HexaPDF version 0.15.5)>>
+      <</Producer(HexaPDF version 0.15.9)>>
       endobj
       2 0 obj
       <</Length 10>>stream

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: hexapdf
 version: !ruby/object:Gem::Version
-  version: 0.15.5
+  version: 0.15.9
 platform: ruby
 authors:
 - Thomas Leitner
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-07-06 00:00:00.000000000 Z
+date: 2021-09-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cmdparse