Reject unclosed DOCTYPE on parsing

Fix ruby#152
makenowjust · Jun 18, 2024 · 4af4265 · 4af4265
1 parent d906ae2
commit 4af4265
Show file tree

Hide file tree

Showing 2 changed files with 68 additions and 11 deletions.
diff --git a/lib/rexml/parsers/treeparser.rb b/lib/rexml/parsers/treeparser.rb
@@ -24,12 +24,20 @@ def parse
             #STDERR.puts "TREEPARSER GOT #{event.inspect}"
             case event[0]
             when :end_document
+              if in_doctype
+                raise ParseException.new("Malformed DOCTYPE: unclosed",
+                                         @parser.source, @parser)
+              end
               unless tag_stack.empty?
                 raise ParseException.new("No close tag for #{@build_context.xpath}",
                                          @parser.source, @parser)
               end
               return
             when :start_element
+              if in_doctype
+                raise ParseException.new("Malformed DOCTYPE: unclosed",
+                                         @parser.source, @parser)
+              end
               tag_stack.push(event[1])
               el = @build_context = @build_context.add_element( event[1] )
               event[2].each do |key, value|
@@ -39,17 +47,19 @@ def parse
               tag_stack.pop
               @build_context = @build_context.parent
             when :text
-              if not in_doctype
-                if @build_context[-1].instance_of? Text
-                  @build_context[-1] << event[1]
-                else
-                  @build_context.add(
-                    Text.new(event[1], @build_context.whitespace, nil, true)
-                  ) unless (
-                    @build_context.ignore_whitespace_nodes and
-                    event[1].strip.size==0
-                  )
-                end
+              if in_doctype
+                raise ParseException.new("Malformed DOCTYPE: unclosed",
+                                         @parser.source, @parser)
+              end
+              if @build_context[-1].instance_of? Text
+                @build_context[-1] << event[1]
+              else
+                @build_context.add(
+                  Text.new(event[1], @build_context.whitespace, nil, true)
+                ) unless (
+                  @build_context.ignore_whitespace_nodes and
+                  event[1].strip.size==0
+                )
               end
             when :comment
               c = Comment.new( event[1] )

diff --git a/test/parse/test_document_type_declaration.rb b/test/parse/test_document_type_declaration.rb
@@ -53,6 +53,53 @@ def test_no_name
       end
     end
 
+    class TestUnclosed < self
+      def test_unclosed_doctype_only
+        exception = assert_raise(REXML::ParseException) do
+          REXML::Document.new(<<~DOCTYPE)
+            <!DOCTYPE foo [
+          DOCTYPE
+        end
+        assert_equal(<<~DETAIL.chomp, exception.to_s)
+          Malformed DOCTYPE: unclosed
+          Line: 1
+          Position: 16
+          Last 80 unconsumed characters:
+
+        DETAIL
+      end
+
+      def test_unclosed_doctype_plus_start_element
+        exception = assert_raise(REXML::ParseException) do
+          REXML::Document.new(<<~DOCTYPE)
+            <!DOCTYPE foo [ <r>
+          DOCTYPE
+        end
+        assert_equal(<<~DETAIL.chomp, exception.to_s)
+          Malformed DOCTYPE: unclosed
+          Line: 1
+          Position: 20
+          Last 80 unconsumed characters:
+          #{' '}
+        DETAIL
+      end
+
+      def test_unclosed_doctype_plus_text
+        exception = assert_raise(REXML::ParseException) do
+          REXML::Document.new(<<~DOCTYPE)
+            <!DOCTYPE foo [ text
+          DOCTYPE
+        end
+        assert_equal(<<~DETAIL.chomp, exception.to_s)
+          Malformed DOCTYPE: unclosed
+          Line: 1
+          Position: 21
+          Last 80 unconsumed characters:
+          
+        DETAIL
+      end
+    end
+
     class TestExternalID < self
       class TestSystem < self
         def test_left_bracket_in_system_literal