From 7e5e2fc262c1d93ebb5a61a310400ff395ac7759 Mon Sep 17 00:00:00 2001
From: TSUYUSATO Kitsune <make.just.on@gmail.com>
Date: Wed, 19 Jun 2024 11:21:43 +0900
Subject: [PATCH] Fix to handle unclosed doctype decls in `BaseParser`

See https://github.com/ruby/rexml/pull/153#discussion_r1645228365
---
 lib/rexml/parsers/baseparser.rb              | 10 +++++++++-
 lib/rexml/parsers/treeparser.rb              | 12 ------------
 test/parse/test_document_type_declaration.rb | 14 ++++++--------
 3 files changed, 15 insertions(+), 21 deletions(-)

diff --git a/lib/rexml/parsers/baseparser.rb b/lib/rexml/parsers/baseparser.rb
index 272d8a6b..5791ab1d 100644
--- a/lib/rexml/parsers/baseparser.rb
+++ b/lib/rexml/parsers/baseparser.rb
@@ -216,7 +216,12 @@ def pull_event
           x, @closed = @closed, nil
           return [ :end_element, x ]
         end
-        return [ :end_document ] if empty?
+        if empty?
+          if @document_status == :in_doctype
+            raise ParseException.new("Malformed DOCTYPE: unclosed", @source)
+          end
+          return [ :end_document ]
+        end
         return @stack.shift if @stack.size > 0
         #STDERR.puts @source.encoding
         #STDERR.puts "BUFFER = #{@source.buffer.inspect}"
@@ -373,6 +378,9 @@ def pull_event
             @document_status = :after_doctype
             return [ :end_doctype ]
           end
+          if @document_status == :in_doctype
+            raise ParseException.new("Malformed DOCTYPE: invalid declaration", @source)
+          end
         end
         if @document_status == :after_doctype
           @source.match(/\s*/um, true)
diff --git a/lib/rexml/parsers/treeparser.rb b/lib/rexml/parsers/treeparser.rb
index 9ddb54db..04dddea9 100644
--- a/lib/rexml/parsers/treeparser.rb
+++ b/lib/rexml/parsers/treeparser.rb
@@ -24,20 +24,12 @@ 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|
@@ -47,10 +39,6 @@ def parse
               tag_stack.pop
               @build_context = @build_context.parent
             when :text
-              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
diff --git a/test/parse/test_document_type_declaration.rb b/test/parse/test_document_type_declaration.rb
index 8a726476..3ca0b536 100644
--- a/test/parse/test_document_type_declaration.rb
+++ b/test/parse/test_document_type_declaration.rb
@@ -56,14 +56,12 @@ def test_no_name
     class TestUnclosed < self
       def test_no_extra_node
         exception = assert_raise(REXML::ParseException) do
-          REXML::Document.new(<<~DOCTYPE)
-            <!DOCTYPE foo [
-          DOCTYPE
+          REXML::Document.new("<!DOCTYPE foo [")
         end
         assert_equal(<<~DETAIL.chomp, exception.to_s)
           Malformed DOCTYPE: unclosed
           Line: 1
-          Position: 16
+          Position: 15
           Last 80 unconsumed characters:
 
         DETAIL
@@ -76,11 +74,11 @@ def test_start_element
           DOCTYPE
         end
         assert_equal(<<~DETAIL.chomp, exception.to_s)
-          Malformed DOCTYPE: unclosed
+          Malformed DOCTYPE: invalid declaration
           Line: 1
           Position: 20
           Last 80 unconsumed characters:
-          #{' '}
+          <r>#{' '}
         DETAIL
       end
 
@@ -91,11 +89,11 @@ def test_text
           DOCTYPE
         end
         assert_equal(<<~DETAIL.chomp, exception.to_s)
-          Malformed DOCTYPE: unclosed
+          Malformed DOCTYPE: invalid declaration
           Line: 1
           Position: 21
           Last 80 unconsumed characters:
-          
+          text#{' '}
         DETAIL
       end
     end