REXML 3.1.9.1

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_2_6@67940 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

11 files changed, 607 insertions, 80 deletions

diff --git a/lib/rexml/doctype.rb b/lib/rexml/doctype.rb
index 757b639639..a4e91529ac 100644
--- a/lib/rexml/doctype.rb
+++ b/lib/rexml/doctype.rb
@@ -7,6 +7,44 @@ require_relative 'attlistdecl'
 require_relative 'xmltokens'
 
 module REXML
+  class ReferenceWriter
+    def initialize(id_type,
+                   public_id_literal,
+                   system_literal,
+                   context=nil)
+      @id_type = id_type
+      @public_id_literal = public_id_literal
+      @system_literal = system_literal
+      if context and context[:prologue_quote] == :apostrophe
+        @default_quote = "'"
+      else
+        @default_quote = "\""
+      end
+    end
+
+    def write(output)
+      output << " #{@id_type}"
+      if @public_id_literal
+        if @public_id_literal.include?("'")
+          quote = "\""
+        else
+          quote = @default_quote
+        end
+        output << " #{quote}#{@public_id_literal}#{quote}"
+      end
+      if @system_literal
+        if @system_literal.include?("'")
+          quote = "\""
+        elsif @system_literal.include?("\"")
+          quote = "'"
+        else
+          quote = @default_quote
+        end
+        output << " #{quote}#{@system_literal}#{quote}"
+      end
+    end
+  end
+
   # Represents an XML DOCTYPE declaration; that is, the contents of <!DOCTYPE
   # ... >.  DOCTYPES can be used to declare the DTD of a document, as well as
   # being used to declare entities used in the document.
@@ -50,6 +88,8 @@ module REXML
         super( parent )
         @name = first.name
         @external_id = first.external_id
+        @long_name = first.instance_variable_get(:@long_name)
+        @uri = first.instance_variable_get(:@uri)
       elsif first.kind_of? Array
         super( parent )
         @name = first[0]
@@ -108,19 +148,17 @@ module REXML
     #   Ignored
     def write( output, indent=0, transitive=false, ie_hack=false )
       f = REXML::Formatters::Default.new
-      c = context
-      if c and c[:prologue_quote] == :apostrophe
-        quote = "'"
-      else
-        quote = "\""
-      end
       indent( output, indent )
       output << START
       output << ' '
       output << @name
-      output << " #{@external_id}" if @external_id
-      output << " #{quote}#{@long_name}#{quote}" if @long_name
-      output << " #{quote}#{@uri}#{quote}" if @uri
+      if @external_id
+        reference_writer = ReferenceWriter.new(@external_id,
+                                               @long_name,
+                                               @uri,
+                                               context)
+        reference_writer.write(output)
+      end
       unless @children.empty?
         output << ' ['
         @children.each { |child|
@@ -259,16 +297,11 @@ module REXML
     end
 
     def to_s
-      c = nil
-      c = parent.context if parent
-      if c and c[:prologue_quote] == :apostrophe
-        quote = "'"
-      else
-        quote = "\""
-      end
-      notation = "<!NOTATION #{@name} #{@middle}"
-      notation << " #{quote}#{@public}#{quote}" if @public
-      notation << " #{quote}#{@system}#{quote}" if @system
+      context = nil
+      context = parent.context if parent
+      notation = "<!NOTATION #{@name}"
+      reference_writer = ReferenceWriter.new(@middle, @public, @system, context)
+      reference_writer.write(notation)
       notation << ">"
       notation
     end
diff --git a/lib/rexml/parsers/baseparser.rb b/lib/rexml/parsers/baseparser.rb
index 39e9ec3fb1..b10cdcd30e 100644
--- a/lib/rexml/parsers/baseparser.rb
+++ b/lib/rexml/parsers/baseparser.rb
@@ -50,7 +50,6 @@ module REXML
 
       DOCTYPE_START = /\A\s*<!DOCTYPE\s/um
       DOCTYPE_END = /\A\s*\]\s*>/um
-      DOCTYPE_PATTERN = /\s*<!DOCTYPE\s+(.*?)(\[|>)/um
       ATTRIBUTE_PATTERN = /\s*(#{QNAME_STR})\s*=\s*(["'])(.*?)\4/um
       COMMENT_START = /\A<!--/u
       COMMENT_PATTERN = /<!--(.*?)-->/um
@@ -61,15 +60,14 @@ module REXML
       XMLDECL_PATTERN = /<\?xml\s+(.*?)\?>/um
       INSTRUCTION_START = /\A<\?/u
       INSTRUCTION_PATTERN = /<\?#{NAME}(\s+.*?)?\?>/um
-      TAG_MATCH = /^<((?>#{QNAME_STR}))/um
-      CLOSE_MATCH = /^\s*<\/(#{QNAME_STR})\s*>/um
+      TAG_MATCH = /\A<((?>#{QNAME_STR}))/um
+      CLOSE_MATCH = /\A\s*<\/(#{QNAME_STR})\s*>/um
 
       VERSION = /\bversion\s*=\s*["'](.*?)['"]/um
       ENCODING = /\bencoding\s*=\s*["'](.*?)['"]/um
       STANDALONE = /\bstandalone\s*=\s*["'](.*?)['"]/um
 
       ENTITY_START = /\A\s*<!ENTITY/
-      IDENTITY = /^([!\*\w\-]+)(\s+#{NCNAME_STR})?(\s+["'](.*?)['"])?(\s+['"](.*?)["'])?/u
       ELEMENTDECL_START = /\A\s*<!ELEMENT/um
       ELEMENTDECL_PATTERN = /\A\s*(<!ELEMENT.*?)>/um
       SYSTEMENTITY = /\A\s*(%.*?;)\s*$/um
@@ -83,9 +81,6 @@ module REXML
       ATTDEF_RE = /#{ATTDEF}/
       ATTLISTDECL_START = /\A\s*<!ATTLIST/um
       ATTLISTDECL_PATTERN = /\A\s*<!ATTLIST\s+#{NAME}(?:#{ATTDEF})*\s*>/um
-      NOTATIONDECL_START = /\A\s*<!NOTATION/um
-      PUBLIC = /\A\s*<!NOTATION\s+(\w[\-\w]*)\s+(PUBLIC)\s+(["'])(.*?)\3(?:\s+(["'])(.*?)\5)?\s*>/um
-      SYSTEM = /\A\s*<!NOTATION\s+(\w[\-\w]*)\s+(SYSTEM)\s+(["'])(.*?)\3\s*>/um
 
       TEXT_PATTERN = /\A([^<]*)/um
 
@@ -103,6 +98,11 @@ module REXML
       GEDECL = "<!ENTITY\\s+#{NAME}\\s+#{ENTITYDEF}\\s*>"
       ENTITYDECL = /\s*(?:#{GEDECL})|(?:#{PEDECL})/um
 
+      NOTATIONDECL_START = /\A\s*<!NOTATION/um
+      EXTERNAL_ID_PUBLIC = /\A\s*PUBLIC\s+#{PUBIDLITERAL}\s+#{SYSTEMLITERAL}\s*/um
+      EXTERNAL_ID_SYSTEM = /\A\s*SYSTEM\s+#{SYSTEMLITERAL}\s*/um
+      PUBLIC_ID = /\A\s*PUBLIC\s+#{PUBIDLITERAL}\s*/um
+
       EREFERENCE = /&(?!#{NAME};)/
 
       DEFAULT_ENTITIES = {
@@ -195,11 +195,9 @@ module REXML
         return [ :end_document ] if empty?
         return @stack.shift if @stack.size > 0
         #STDERR.puts @source.encoding
-        @source.read if @source.buffer.size<2
         #STDERR.puts "BUFFER = #{@source.buffer.inspect}"
         if @document_status == nil
-          #@source.consume( /^\s*/um )
-          word = @source.match( /^((?:\s+)|(?:<[^>]*>))/um )
+          word = @source.match( /\A((?:\s+)|(?:<[^>]*>))/um )
           word = word[1] unless word.nil?
           #STDERR.puts "WORD = #{word.inspect}"
           case word
@@ -224,38 +222,49 @@ module REXML
           when INSTRUCTION_START
             return process_instruction
           when DOCTYPE_START
-            md = @source.match( DOCTYPE_PATTERN, true )
+            base_error_message = "Malformed DOCTYPE"
+            @source.match(DOCTYPE_START, true)
             @nsstack.unshift(curr_ns=Set.new)
-            identity = md[1]
-            close = md[2]
-            identity =~ IDENTITY
-            name = $1
-            raise REXML::ParseException.new("DOCTYPE is missing a name") if name.nil?
-            pub_sys = $2.nil? ? nil : $2.strip
-            long_name = $4.nil? ? nil : $4.strip
-            uri = $6.nil? ? nil : $6.strip
-            args = [ :start_doctype, name, pub_sys, long_name, uri ]
-            if close == ">"
+            name = parse_name(base_error_message)
+            if @source.match(/\A\s*\[/um, true)
+              id = [nil, nil, nil]
+              @document_status = :in_doctype
+            elsif @source.match(/\A\s*>/um, true)
+              id = [nil, nil, nil]
               @document_status = :after_doctype
-              @source.read if @source.buffer.size<2
-              md = @source.match(/^\s*/um, true)
-              @stack << [ :end_doctype ]
             else
-              @document_status = :in_doctype
+              id = parse_id(base_error_message,
+                            accept_external_id: true,
+                            accept_public_id: false)
+              if id[0] == "SYSTEM"
+                # For backward compatibility
+                id[1], id[2] = id[2], nil
+              end
+              if @source.match(/\A\s*\[/um, true)
+                @document_status = :in_doctype
+              elsif @source.match(/\A\s*>/um, true)
+                @document_status = :after_doctype
+              else
+                message = "#{base_error_message}: garbage after external ID"
+                raise REXML::ParseException.new(message, @source)
+              end
+            end
+            args = [:start_doctype, name, *id]
+            if @document_status == :after_doctype
+              @source.match(/\A\s*/um, true)
+              @stack << [ :end_doctype ]
             end
             return args
-          when /^\s+/
+          when /\A\s+/
           else
             @document_status = :after_doctype
-            @source.read if @source.buffer.size<2
-            md = @source.match(/\s*/um, true)
             if @source.encoding == "UTF-8"
               @source.buffer.force_encoding(::Encoding::UTF_8)
             end
           end
         end
         if @document_status == :in_doctype
-          md = @source.match(/\s*(.*?>)/um)
+          md = @source.match(/\A\s*(.*?>)/um)
           case md[1]
           when SYSTEMENTITY
             match = @source.match( SYSTEMENTITY, true )[1]
@@ -312,24 +321,35 @@ module REXML
             end
             return [ :attlistdecl, element, pairs, contents ]
           when NOTATIONDECL_START
-            md = nil
-            if @source.match( PUBLIC )
-              md = @source.match( PUBLIC, true )
-              vals = [md[1],md[2],md[4],md[6]]
-            elsif @source.match( SYSTEM )
-              md = @source.match( SYSTEM, true )
-              vals = [md[1],md[2],nil,md[4]]
-            else
-              raise REXML::ParseException.new( "error parsing notation: no matching pattern", @source )
+            base_error_message = "Malformed notation declaration"
+            unless @source.match(/\A\s*<!NOTATION\s+/um, true)
+              if @source.match(/\A\s*<!NOTATION\s*>/um)
+                message = "#{base_error_message}: name is missing"
+              else
+                message = "#{base_error_message}: invalid declaration name"
+              end
+              raise REXML::ParseException.new(message, @source)
             end
-            return [ :notationdecl, *vals ]
+            name = parse_name(base_error_message)
+            id = parse_id(base_error_message,
+                          accept_external_id: true,
+                          accept_public_id: true)
+            unless @source.match(/\A\s*>/um, true)
+              message = "#{base_error_message}: garbage before end >"
+              raise REXML::ParseException.new(message, @source)
+            end
+            return [:notationdecl, name, *id]
           when DOCTYPE_END
             @document_status = :after_doctype
             @source.match( DOCTYPE_END, true )
             return [ :end_doctype ]
           end
         end
+        if @document_status == :after_doctype
+          @source.match(/\A\s*/um, true)
+        end
         begin
+          @source.read if @source.buffer.size<2
           if @source.buffer[0] == ?<
             if @source.buffer[1] == ?/
               @nsstack.shift
@@ -368,6 +388,7 @@ module REXML
               unless md
                 raise REXML::ParseException.new("malformed XML: missing tag start", @source)
               end
+              @document_status = :in_element
               prefixes = Set.new
               prefixes << md[2] if md[2]
               @nsstack.unshift(curr_ns=Set.new)
@@ -473,6 +494,85 @@ module REXML
         true
       end
 
+      def parse_name(base_error_message)
+        md = @source.match(/\A\s*#{NAME}/um, true)
+        unless md
+          if @source.match(/\A\s*\S/um)
+            message = "#{base_error_message}: invalid name"
+          else
+            message = "#{base_error_message}: name is missing"
+          end
+          raise REXML::ParseException.new(message, @source)
+        end
+        md[1]
+      end
+
+      def parse_id(base_error_message,
+                   accept_external_id:,
+                   accept_public_id:)
+        if accept_external_id and (md = @source.match(EXTERNAL_ID_PUBLIC, true))
+          pubid = system = nil
+          pubid_literal = md[1]
+          pubid = pubid_literal[1..-2] if pubid_literal # Remove quote
+          system_literal = md[2]
+          system = system_literal[1..-2] if system_literal # Remove quote
+          ["PUBLIC", pubid, system]
+        elsif accept_public_id and (md = @source.match(PUBLIC_ID, true))
+          pubid = system = nil
+          pubid_literal = md[1]
+          pubid = pubid_literal[1..-2] if pubid_literal # Remove quote
+          ["PUBLIC", pubid, nil]
+        elsif accept_external_id and (md = @source.match(EXTERNAL_ID_SYSTEM, true))
+          system = nil
+          system_literal = md[1]
+          system = system_literal[1..-2] if system_literal # Remove quote
+          ["SYSTEM", nil, system]
+        else
+          details = parse_id_invalid_details(accept_external_id: accept_external_id,
+                                             accept_public_id: accept_public_id)
+          message = "#{base_error_message}: #{details}"
+          raise REXML::ParseException.new(message, @source)
+        end
+      end
+
+      def parse_id_invalid_details(accept_external_id:,
+                                   accept_public_id:)
+        public = /\A\s*PUBLIC/um
+        system = /\A\s*SYSTEM/um
+        if (accept_external_id or accept_public_id) and @source.match(/#{public}/um)
+          if @source.match(/#{public}(?:\s+[^'"]|\s*[\[>])/um)
+            return "public ID literal is missing"
+          end
+          unless @source.match(/#{public}\s+#{PUBIDLITERAL}/um)
+            return "invalid public ID literal"
+          end
+          if accept_public_id
+            if @source.match(/#{public}\s+#{PUBIDLITERAL}\s+[^'"]/um)
+              return "system ID literal is missing"
+            end
+            unless @source.match(/#{public}\s+#{PUBIDLITERAL}\s+#{SYSTEMLITERAL}/um)
+              return "invalid system literal"
+            end
+            "garbage after system literal"
+          else
+            "garbage after public ID literal"
+          end
+        elsif accept_external_id and @source.match(/#{system}/um)
+          if @source.match(/#{system}(?:\s+[^'"]|\s*[\[>])/um)
+            return "system literal is missing"
+          end
+          unless @source.match(/#{system}\s+#{SYSTEMLITERAL}/um)
+            return "invalid system literal"
+          end
+          "garbage after system literal"
+        else
+          unless @source.match(/\A\s*(?:PUBLIC|SYSTEM)\s/um)
+            return "invalid ID type"
+          end
+          "ID type is missing"
+        end
+      end
+
       def process_instruction
         match_data = @source.match(INSTRUCTION_PATTERN, true)
         unless match_data
diff --git a/lib/rexml/rexml.rb b/lib/rexml/rexml.rb
index 7851c78bf5..627dff4768 100644
--- a/lib/rexml/rexml.rb
+++ b/lib/rexml/rexml.rb
@@ -24,7 +24,7 @@
 module REXML
   COPYRIGHT = "Copyright © 2001-2008 Sean Russell <ser@germane-software.com>"
   DATE = "2008/019"
-  VERSION = "3.1.9"
+  VERSION = "3.1.9.1"
   REVISION = ""
 
   Copyright = COPYRIGHT
diff --git a/test/rexml/parse/test_document_type_declaration.rb b/test/rexml/parse/test_document_type_declaration.rb
index 80f70888fb..55713909e7 100644
--- a/test/rexml/parse/test_document_type_declaration.rb
+++ b/test/rexml/parse/test_document_type_declaration.rb
@@ -5,17 +5,187 @@ require "rexml/document"
 module REXMLTests
   class TestParseDocumentTypeDeclaration < Test::Unit::TestCase
     private
-    def xml(internal_subset)
-      <<-XML
-<!DOCTYPE r SYSTEM "urn:x-rexml:test" [
-#{internal_subset}
-]>
+    def parse(doctype)
+      REXML::Document.new(<<-XML).doctype
+#{doctype}
 <r/>
       XML
     end
 
-    def parse(internal_subset)
-      REXML::Document.new(xml(internal_subset)).doctype
+    class TestName < self
+      def test_valid
+        doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r>
+        DOCTYPE
+        assert_equal("r", doctype.name)
+      end
+
+      def test_garbage_plus_before_name_at_line_start
+        exception = assert_raise(REXML::ParseException) do
+          parse(<<-DOCTYPE)
+<!DOCTYPE +
+r SYSTEM "urn:x-rexml:test" [
+]>
+          DOCTYPE
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed DOCTYPE: invalid name
+Line: 5
+Position: 51
+Last 80 unconsumed characters:
++ r SYSTEM "urn:x-rexml:test" [ ]>  <r/> 
+        DETAIL
+      end
+    end
+
+    class TestExternalID < self
+      class TestSystem < self
+        def test_left_bracket_in_system_literal
+          doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r SYSTEM "urn:x-rexml:[test" [
+]>
+          DOCTYPE
+          assert_equal([
+                         "r",
+                         "SYSTEM",
+                         nil,
+                         "urn:x-rexml:[test",
+                       ],
+                       [
+                         doctype.name,
+                         doctype.external_id,
+                         doctype.public,
+                         doctype.system,
+                       ])
+        end
+
+        def test_greater_than_in_system_literal
+          doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r SYSTEM "urn:x-rexml:>test" [
+]>
+          DOCTYPE
+          assert_equal([
+                         "r",
+                         "SYSTEM",
+                         nil,
+                         "urn:x-rexml:>test",
+                       ],
+                       [
+                         doctype.name,
+                         doctype.external_id,
+                         doctype.public,
+                         doctype.system,
+                       ])
+        end
+
+        def test_no_literal
+          exception = assert_raise(REXML::ParseException) do
+            parse(<<-DOCTYPE)
+<!DOCTYPE r SYSTEM>
+            DOCTYPE
+          end
+          assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed DOCTYPE: system literal is missing
+Line: 3
+Position: 26
+Last 80 unconsumed characters:
+ SYSTEM>  <r/> 
+          DETAIL
+        end
+
+        def test_garbage_after_literal
+          exception = assert_raise(REXML::ParseException) do
+            parse(<<-DOCTYPE)
+<!DOCTYPE r SYSTEM 'r.dtd'x'>
+            DOCTYPE
+          end
+          assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed DOCTYPE: garbage after external ID
+Line: 3
+Position: 36
+Last 80 unconsumed characters:
+x'>  <r/> 
+          DETAIL
+        end
+
+        def test_single_quote
+          doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r SYSTEM 'r".dtd'>
+          DOCTYPE
+          assert_equal("r\".dtd", doctype.system)
+        end
+
+        def test_double_quote
+          doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r SYSTEM "r'.dtd">
+          DOCTYPE
+          assert_equal("r'.dtd", doctype.system)
+        end
+      end
+
+      class TestPublic < self
+        class TestPublicIDLiteral < self
+          def test_content_double_quote
+            exception = assert_raise(REXML::ParseException) do
+              parse(<<-DOCTYPE)
+<!DOCTYPE r PUBLIC 'double quote " is invalid' "r.dtd">
+              DOCTYPE
+            end
+            assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed DOCTYPE: invalid public ID literal
+Line: 3
+Position: 62
+Last 80 unconsumed characters:
+ PUBLIC 'double quote " is invalid' "r.dtd">  <r/> 
+            DETAIL
+          end
+
+          def test_single_quote
+            doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r PUBLIC 'public-id-literal' "r.dtd">
+            DOCTYPE
+            assert_equal("public-id-literal", doctype.public)
+          end
+
+          def test_double_quote
+            doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r PUBLIC "public'-id-literal" "r.dtd">
+            DOCTYPE
+            assert_equal("public'-id-literal", doctype.public)
+          end
+        end
+
+        class TestSystemLiteral < self
+          def test_garbage_after_literal
+            exception = assert_raise(REXML::ParseException) do
+              parse(<<-DOCTYPE)
+<!DOCTYPE r PUBLIC 'public-id-literal' 'system-literal'x'>
+              DOCTYPE
+            end
+            assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed DOCTYPE: garbage after external ID
+Line: 3
+Position: 65
+Last 80 unconsumed characters:
+x'>  <r/> 
+           DETAIL
+          end
+
+          def test_single_quote
+            doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r PUBLIC "public-id-literal" 'system"-literal'>
+            DOCTYPE
+            assert_equal("system\"-literal", doctype.system)
+          end
+
+          def test_double_quote
+            doctype = parse(<<-DOCTYPE)
+<!DOCTYPE r PUBLIC "public-id-literal" "system'-literal">
+            DOCTYPE
+            assert_equal("system'-literal", doctype.system)
+          end
+        end
+      end
     end
 
     class TestMixed < self
@@ -45,6 +215,15 @@ module REXMLTests
         assert_equal([REXML::NotationDecl, REXML::AttlistDecl],
                      doctype.children.collect(&:class))
       end
+
+      private
+      def parse(internal_subset)
+        super(<<-DOCTYPE)
+<!DOCTYPE r SYSTEM "urn:x-rexml:test" [
+#{internal_subset}
+]>
+        DOCTYPE
+      end
     end
   end
 end
diff --git a/test/rexml/parse/test_element.rb b/test/rexml/parse/test_element.rb
index aad915fe7b..7206fe595a 100644
--- a/test/rexml/parse/test_element.rb
+++ b/test/rexml/parse/test_element.rb
@@ -33,6 +33,32 @@ Last 80 unconsumed characters:
 
         DETAIL
       end
+
+      def test_garbage_less_than_before_root_element_at_line_start
+        exception = assert_raise(REXML::ParseException) do
+          parse("<\n<x/>")
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+malformed XML: missing tag start
+Line: 2
+Position: 6
+Last 80 unconsumed characters:
+< <x/>
+        DETAIL
+      end
+
+      def test_garbage_less_than_slash_before_end_tag_at_line_start
+        exception = assert_raise(REXML::ParseException) do
+          parse("<x></\n</x>")
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Missing end tag for 'x'
+Line: 2
+Position: 10
+Last 80 unconsumed characters:
+</ </x>
+        DETAIL
+      end
     end
   end
 end
diff --git a/test/rexml/parse/test_notation_declaration.rb b/test/rexml/parse/test_notation_declaration.rb
index 0d29f0d81f..19a0536d0a 100644
--- a/test/rexml/parse/test_notation_declaration.rb
+++ b/test/rexml/parse/test_notation_declaration.rb
@@ -23,10 +23,100 @@ module REXMLTests
         doctype = parse("<!NOTATION name PUBLIC 'urn:public-id'>")
         assert_equal("name", doctype.notation("name").name)
       end
+
+      def test_no_name
+        exception = assert_raise(REXML::ParseException) do
+          parse(<<-INTERNAL_SUBSET)
+<!NOTATION>
+          INTERNAL_SUBSET
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: name is missing
+Line: 5
+Position: 72
+Last 80 unconsumed characters:
+ <!NOTATION>  ]> <r/> 
+        DETAIL
+      end
+
+      def test_invalid_name
+        exception = assert_raise(REXML::ParseException) do
+          parse(<<-INTERNAL_SUBSET)
+<!NOTATION '>
+          INTERNAL_SUBSET
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: invalid name
+Line: 5
+Position: 74
+Last 80 unconsumed characters:
+'>  ]> <r/> 
+        DETAIL
+      end
+
+      def test_no_id_type
+        exception = assert_raise(REXML::ParseException) do
+          parse(<<-INTERNAL_SUBSET)
+<!NOTATION name>
+          INTERNAL_SUBSET
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: invalid ID type
+Line: 5
+Position: 77
+Last 80 unconsumed characters:
+>  ]> <r/> 
+        DETAIL
+      end
+
+      def test_invalid_id_type
+        exception = assert_raise(REXML::ParseException) do
+          parse(<<-INTERNAL_SUBSET)
+<!NOTATION name INVALID>
+          INTERNAL_SUBSET
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: invalid ID type
+Line: 5
+Position: 85
+Last 80 unconsumed characters:
+ INVALID>  ]> <r/> 
+        DETAIL
+      end
     end
 
     class TestExternalID < self
       class TestSystem < self
+        def test_no_literal
+          exception = assert_raise(REXML::ParseException) do
+            parse(<<-INTERNAL_SUBSET)
+<!NOTATION name SYSTEM>
+            INTERNAL_SUBSET
+          end
+          assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: system literal is missing
+Line: 5
+Position: 84
+Last 80 unconsumed characters:
+ SYSTEM>  ]> <r/> 
+          DETAIL
+        end
+
+        def test_garbage_after_literal
+          exception = assert_raise(REXML::ParseException) do
+            parse(<<-INTERNAL_SUBSET)
+<!NOTATION name SYSTEM 'system-literal'x'>
+            INTERNAL_SUBSET
+          end
+          assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: garbage before end >
+Line: 5
+Position: 103
+Last 80 unconsumed characters:
+x'>  ]> <r/> 
+          DETAIL
+        end
+
         def test_single_quote
           doctype = parse(<<-INTERNAL_SUBSET)
 <!NOTATION name SYSTEM 'system-literal'>
@@ -44,6 +134,21 @@ module REXMLTests
 
       class TestPublic < self
         class TestPublicIDLiteral < self
+          def test_content_double_quote
+            exception = assert_raise(REXML::ParseException) do
+              parse(<<-INTERNAL_SUBSET)
+<!NOTATION name PUBLIC 'double quote " is invalid' "system-literal">
+              INTERNAL_SUBSET
+            end
+            assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: invalid public ID literal
+Line: 5
+Position: 129
+Last 80 unconsumed characters:
+ PUBLIC 'double quote " is invalid' "system-literal">  ]> <r/> 
+            DETAIL
+          end
+
           def test_single_quote
             doctype = parse(<<-INTERNAL_SUBSET)
 <!NOTATION name PUBLIC 'public-id-literal' "system-literal">
@@ -60,6 +165,21 @@ module REXMLTests
         end
 
         class TestSystemLiteral < self
+          def test_garbage_after_literal
+            exception = assert_raise(REXML::ParseException) do
+              parse(<<-INTERNAL_SUBSET)
+<!NOTATION name PUBLIC 'public-id-literal' 'system-literal'x'>
+              INTERNAL_SUBSET
+            end
+            assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: garbage before end >
+Line: 5
+Position: 123
+Last 80 unconsumed characters:
+x'>  ]> <r/> 
+           DETAIL
+          end
+
           def test_single_quote
             doctype = parse(<<-INTERNAL_SUBSET)
 <!NOTATION name PUBLIC "public-id-literal" 'system-literal'>
@@ -96,5 +216,66 @@ module REXMLTests
         end
       end
     end
+
+    class TestPublicID < self
+      def test_no_literal
+        exception = assert_raise(REXML::ParseException) do
+          parse(<<-INTERNAL_SUBSET)
+<!NOTATION name PUBLIC>
+          INTERNAL_SUBSET
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: public ID literal is missing
+Line: 5
+Position: 84
+Last 80 unconsumed characters:
+ PUBLIC>  ]> <r/> 
+        DETAIL
+      end
+
+      def test_literal_content_double_quote
+        exception = assert_raise(REXML::ParseException) do
+          parse(<<-INTERNAL_SUBSET)
+<!NOTATION name PUBLIC 'double quote " is invalid in PubidLiteral'>
+          INTERNAL_SUBSET
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: invalid public ID literal
+Line: 5
+Position: 128
+Last 80 unconsumed characters:
+ PUBLIC 'double quote \" is invalid in PubidLiteral'>  ]> <r/> 
+        DETAIL
+      end
+
+      def test_garbage_after_literal
+        exception = assert_raise(REXML::ParseException) do
+          parse(<<-INTERNAL_SUBSET)
+<!NOTATION name PUBLIC 'public-id-literal'x'>
+          INTERNAL_SUBSET
+        end
+        assert_equal(<<-DETAIL.chomp, exception.to_s)
+Malformed notation declaration: garbage before end >
+Line: 5
+Position: 106
+Last 80 unconsumed characters:
+x'>  ]> <r/> 
+        DETAIL
+      end
+
+      def test_literal_single_quote
+        doctype = parse(<<-INTERNAL_SUBSET)
+<!NOTATION name PUBLIC 'public-id-literal'>
+        INTERNAL_SUBSET
+        assert_equal("public-id-literal", doctype.notation("name").public)
+      end
+
+      def test_literal_double_quote
+        doctype = parse(<<-INTERNAL_SUBSET)
+<!NOTATION name PUBLIC "public-id-literal">
+        INTERNAL_SUBSET
+        assert_equal("public-id-literal", doctype.notation("name").public)
+      end
+    end
   end
 end
diff --git a/test/rexml/parse/test_processing_instruction.rb b/test/rexml/parse/test_processing_instruction.rb
index a23513fc6e..f0c0c24e67 100644
--- a/test/rexml/parse/test_processing_instruction.rb
+++ b/test/rexml/parse/test_processing_instruction.rb
@@ -20,6 +20,25 @@ Last 80 unconsumed characters:
 <??>
         DETAIL
       end
+
+      def test_garbage_text
+        # TODO: This should be parse error.
+        # Create test/parse/test_document.rb or something and move this to it.
+        doc = parse(<<-XML)
+x<?x y
+<!--?><?x -->?>
+<r/>
+        XML
+        pi = doc.children[1]
+        assert_equal([
+                       "x",
+                       "y\n<!--",
+                     ],
+                     [
+                       pi.target,
+                       pi.content,
+                     ])
+      end
     end
   end
 end
diff --git a/test/rexml/parser/test_ultra_light.rb b/test/rexml/parser/test_ultra_light.rb
index 8f4a3980d5..44fd1d1ec0 100644
--- a/test/rexml/parser/test_ultra_light.rb
+++ b/test/rexml/parser/test_ultra_light.rb
@@ -16,7 +16,6 @@ class TestUltraLightParser < Test::Unit::TestCase
                        nil,
                        [:entitydecl, "name", "value"]
                      ],
-                     [:text, "\n"],
                      [:start_element, :parent, "root", {}],
                      [:text, "\n"],
                    ],
diff --git a/test/rexml/test_core.rb b/test/rexml/test_core.rb
index 46036d7f12..ae528d75d3 100644
--- a/test/rexml/test_core.rb
+++ b/test/rexml/test_core.rb
@@ -1,4 +1,4 @@
-# coding: binary
+# coding: utf-8
 # frozen_string_literal: false
 
 require_relative "rexml_test_utils"
@@ -995,7 +995,7 @@ EOL
       document.write(s)
 
       ## XML Doctype
-      str = '<!DOCTYPE foo "bar">'
+      str = '<!DOCTYPE foo SYSTEM "bar">'
       source  = REXML::Source.new(str)
       doctype = REXML::DocType.new(source)
       document.add(doctype)
diff --git a/test/rexml/test_doctype.rb b/test/rexml/test_doctype.rb
index 7f42669170..915717de8e 100644
--- a/test/rexml/test_doctype.rb
+++ b/test/rexml/test_doctype.rb
@@ -18,12 +18,6 @@ module REXMLTests
       @doc_type_system = REXML::Document.new(xml_system).doctype
 
       @pubid = "TEST_ID"
-      xml_public = <<-XML
-      <!DOCTYPE root PUBLIC "#{@pubid}">
-      <root/>
-      XML
-      @doc_type_public = REXML::Document.new(xml_public).doctype
-
       xml_public_system = <<-XML
       <!DOCTYPE root PUBLIC "#{@pubid}" "#{@sysid}">
       <root/>
@@ -35,11 +29,9 @@ module REXMLTests
       assert_equal([
                      nil,
                      @pubid,
-                     @pubid,
                    ],
                    [
                      @doc_type_system.public,
-                     @doc_type_public.public,
                      @doc_type_public_system.public,
                    ])
     end
@@ -58,12 +50,10 @@ module REXMLTests
     def test_system
       assert_equal([
                      @sysid,
-                     nil,
                      @sysid,
                    ],
                    [
                      @doc_type_system.system,
-                     @doc_type_public.system,
                      @doc_type_public_system.system,
                    ])
     end
diff --git a/version.h b/version.h
index c435e392b5..1c491eb59c 100644
--- a/version.h
+++ b/version.h
@@ -1,6 +1,6 @@
 #define RUBY_VERSION "2.6.7"
 #define RUBY_RELEASE_DATE "2021-04-05"
-#define RUBY_PATCHLEVEL 196
+#define RUBY_PATCHLEVEL 197
 
 #define RUBY_RELEASE_YEAR 2021
 #define RUBY_RELEASE_MONTH 4