hexapdf 0.15.6 → 0.15.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: bcffb10babdbf723a478ea40721d9515222adf8d12ba9bd799f7b95fa66bc408
-  data.tar.gz: d5929900ab1b010a39964edc366ea223ef8a2ee9bcd1e9a5873874b4d5a8ecc5
+  metadata.gz: 1385aca5e91916034a5494142b4c88e51de46d2d13b79ddaed9494c74808793a
+  data.tar.gz: 4fee33d3c96e74c00565ac6211901f39c0242cd2e0926f0760be7bfb18fe7f12
 SHA512:
-  metadata.gz: d12bbd49204c28675d399477ce0249140bf6ead3fe9332541128802f0edc3ebb2b187752b464b9acd3b71ca4ce6cb5cba33caf19282bd5a15020cec3c6e20297
-  data.tar.gz: 8df3586c8069db615bf317f22b28069f4bfd1395d31c285bb5c4a63b1b46ec60b088082a5d461adafd590718df6307734361e1e61fe3210009d158fc95c558be
+  metadata.gz: 3fa1454ec6821500c1f94981ad17efcbf36f125a29870a62ad0d626fe65cd35bb7ef6426021daba3b2554dcbd20f1ce6efc4d93c1d4d8b5303d6063eb27804fb
+  data.tar.gz: 8f2c3de849fed113c6f4fe7494312a202a872f7364052b584a38352315a4a358f135beea8dd951c29d2dbd3b842c4eefe3892ed3d9bb3c24e6875cdbb0c59123

data/CHANGELOG.md

@@ -1,3 +1,11 @@
+## 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

data/lib/hexapdf/parser.rb

@@ -447,6 +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 = '0.15.6'
+  VERSION = '0.15.7'
 
 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

@@ -619,7 +619,12 @@ 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 0.15.6)>>
+      <</Producer(HexaPDF version 0.15.7)>>
       endobj
       xref
       3 1
@@ -72,7 +72,7 @@ describe HexaPDF::Writer do
       141
       %%EOF
       6 0 obj
-      <</Producer(HexaPDF version 0.15.6)>>
+      <</Producer(HexaPDF version 0.15.7)>>
       endobj
       2 0 obj
       <</Length 10>>stream

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: hexapdf
 version: !ruby/object:Gem::Version
-  version: 0.15.6
+  version: 0.15.7
 platform: ruby
 authors:
 - Thomas Leitner
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-07-16 00:00:00.000000000 Z
+date: 2021-07-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cmdparse