rexml 3.2.3 → 3.3.8

data/lib/rexml/parsers/pullparser.rb

@@ -47,6 +47,18 @@ module REXML
         @listeners << listener
       end
 
+      def entity_expansion_count
+        @parser.entity_expansion_count
+      end
+
+      def entity_expansion_limit=( limit )
+        @parser.entity_expansion_limit = limit
+      end
+
+      def entity_expansion_text_limit=( limit )
+        @parser.entity_expansion_text_limit = limit
+      end
+
       def each
         while has_next?
           yield self.pull

data/lib/rexml/parsers/sax2parser.rb

@@ -22,6 +22,18 @@ module REXML
         @parser.source
       end
 
+      def entity_expansion_count
+        @parser.entity_expansion_count
+      end
+
+      def entity_expansion_limit=( limit )
+        @parser.entity_expansion_limit = limit
+      end
+
+      def entity_expansion_text_limit=( limit )
+        @parser.entity_expansion_text_limit = limit
+      end
+
       def add_listener( listener )
         @parser.add_listener( listener )
       end
@@ -157,25 +169,8 @@ module REXML
               end
             end
           when :text
-            #normalized = @parser.normalize( event[1] )
-            #handle( :characters, normalized )
-            copy = event[1].clone
-
-            esub = proc { |match|
-              if @entities.has_key?($1)
-                @entities[$1].gsub(Text::REFERENCE, &esub)
-              else
-                match
-              end
-            }
-
-            copy.gsub!( Text::REFERENCE, &esub )
-            copy.gsub!( Text::NUMERICENTITY ) {|m|
-              m=$1
-              m = "0#{m}" if m[0] == ?x
-              [Integer(m)].pack('U*')
-            }
-            handle( :characters, copy )
+            unnormalized = @parser.unnormalize( event[1], @entities )
+            handle( :characters, unnormalized )
           when :entitydecl
             handle_entitydecl( event )
           when :processing_instruction, :comment, :attlistdecl,
@@ -264,6 +259,8 @@ module REXML
       end
 
       def get_namespace( prefix )
+        return nil if @namespace_stack.empty?
+
         uris = (@namespace_stack.find_all { |ns| not ns[prefix].nil? }) ||
           (@namespace_stack.find { |ns| not ns[nil].nil? })
         uris[-1][prefix] unless uris.nil? or 0 == uris.size

data/lib/rexml/parsers/streamparser.rb

@@ -7,37 +7,42 @@ module REXML
       def initialize source, listener
         @listener = listener
         @parser = BaseParser.new( source )
-        @tag_stack = []
+        @entities = {}
       end
 
       def add_listener( listener )
         @parser.add_listener( listener )
       end
 
+      def entity_expansion_count
+        @parser.entity_expansion_count
+      end
+
+      def entity_expansion_limit=( limit )
+        @parser.entity_expansion_limit = limit
+      end
+
+      def entity_expansion_text_limit=( limit )
+        @parser.entity_expansion_text_limit = limit
+      end
+
       def parse
         # entity string
         while true
           event = @parser.pull
           case event[0]
           when :end_document
-            unless @tag_stack.empty?
-              tag_path = "/" + @tag_stack.join("/")
-              raise ParseException.new("Missing end tag for '#{tag_path}'",
-                                       @parser.source)
-            end
             return
           when :start_element
-            @tag_stack << event[1]
             attrs = event[2].each do |n, v|
               event[2][n] = @parser.unnormalize( v )
             end
             @listener.tag_start( event[1], attrs )
           when :end_element
             @listener.tag_end( event[1] )
-            @tag_stack.pop
           when :text
-            normalized = @parser.unnormalize( event[1] )
-            @listener.text( normalized )
+            unnormalized = @parser.unnormalize( event[1], @entities )
+            @listener.text( unnormalized )
           when :processing_instruction
             @listener.instruction( *event[1,2] )
           when :start_doctype
@@ -48,6 +53,7 @@ module REXML
           when :comment, :attlistdecl, :cdata, :xmldecl, :elementdecl
             @listener.send( event[0].to_s, *event[1..-1] )
           when :entitydecl, :notationdecl
+            @entities[ event[1] ] = event[2] if event.size == 3
             @listener.send( event[0].to_s, event[1..-1] )
           when :externalentity
             entity_reference = event[1]

data/lib/rexml/parsers/treeparser.rb

@@ -15,8 +15,6 @@ module REXML
       end
 
       def parse
-        tag_stack = []
-        in_doctype = false
         entities = nil
         begin
           while true
@@ -24,32 +22,24 @@ module REXML
             #STDERR.puts "TREEPARSER GOT #{event.inspect}"
             case event[0]
             when :end_document
-              unless tag_stack.empty?
-                raise ParseException.new("No close tag for #{@build_context.xpath}",
-                                         @parser.source, @parser)
-              end
               return
             when :start_element
-              tag_stack.push(event[1])
               el = @build_context = @build_context.add_element( event[1] )
               event[2].each do |key, value|
                 el.attributes[key]=Attribute.new(key,value,self)
               end
             when :end_element
-              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 @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] )
@@ -60,14 +50,12 @@ module REXML
             when :processing_instruction
               @build_context.add( Instruction.new( event[1], event[2] ) )
             when :end_doctype
-              in_doctype = false
               entities.each { |k,v| entities[k] = @build_context.entities[k].value }
               @build_context = @build_context.parent
             when :start_doctype
               doctype = DocType.new( event[1..-1], @build_context )
               @build_context = doctype
               entities = {}
-              in_doctype = true
             when :attlistdecl
               n = AttlistDecl.new( event[1..-1] )
               @build_context.add( n )

data/lib/rexml/parsers/xpathparser.rb

@@ -1,4 +1,5 @@
 # frozen_string_literal: false
+
 require_relative '../namespace'
 require_relative '../xmltokens'
 
@@ -22,7 +23,13 @@ module REXML
         path.gsub!(/([\(\[])\s+/, '\1') # Strip ignorable spaces
         path.gsub!( /\s+([\]\)])/, '\1')
         parsed = []
-        OrExpr(path, parsed)
+        rest = OrExpr(path, parsed)
+        if rest
+          unless rest.strip.empty?
+            raise ParseException.new("Garbage component exists at the end: " +
+                                     "<#{rest}>: <#{path}>")
+          end
+        end
         parsed
       end
 
@@ -32,108 +39,143 @@ module REXML
         parsed
       end
 
-      def abbreviate( path )
-        path = path.kind_of?(String) ? parse( path ) : path
-        string = ""
-        document = false
-        while path.size > 0
-          op = path.shift
+      def abbreviate(path_or_parsed)
+        if path_or_parsed.kind_of?(String)
+          parsed = parse(path_or_parsed)
+        else
+          parsed = path_or_parsed
+        end
+        components = []
+        component = nil
+        while parsed.size > 0
+          op = parsed.shift
           case op
           when :node
+            component << "node()"
           when :attribute
-            string << "/" if string.size > 0
-            string << "@"
+            component = "@"
+            components << component
           when :child
-            string << "/" if string.size > 0
+            component = ""
+            components << component
           when :descendant_or_self
-            string << "/"
+            next_op = parsed[0]
+            if next_op == :node
+              parsed.shift
+              component = ""
+              components << component
+            else
+              component = "descendant-or-self::"
+              components << component
+            end
           when :self
-            string << "."
+            next_op = parsed[0]
+            if next_op == :node
+              parsed.shift
+              components << "."
+            else
+              component = "self::"
+              components << component
+            end
           when :parent
-            string << ".."
+            next_op = parsed[0]
+            if next_op == :node
+              parsed.shift
+              components << ".."
+            else
+              component = "parent::"
+              components << component
+            end
           when :any
-            string << "*"
+            component << "*"
           when :text
-            string << "text()"
+            component << "text()"
           when :following, :following_sibling,
                 :ancestor, :ancestor_or_self, :descendant,
                 :namespace, :preceding, :preceding_sibling
-            string << "/" unless string.size == 0
-            string << op.to_s.tr("_", "-")
-            string << "::"
+            component = op.to_s.tr("_", "-") << "::"
+            components << component
           when :qname
-            prefix = path.shift
-            name = path.shift
-            string << prefix+":" if prefix.size > 0
-            string << name
+            prefix = parsed.shift
+            name = parsed.shift
+            component << prefix+":" if prefix.size > 0
+            component << name
           when :predicate
-            string << '['
-            string << predicate_to_string( path.shift ) {|x| abbreviate( x ) }
-            string << ']'
+            component << '['
+            component << predicate_to_path(parsed.shift) {|x| abbreviate(x)}
+            component << ']'
           when :document
-            document = true
+            components << ""
           when :function
-            string << path.shift
-            string << "( "
-            string << predicate_to_string( path.shift[0] ) {|x| abbreviate( x )}
-            string << " )"
+            component << parsed.shift
+            component << "( "
+            component << predicate_to_path(parsed.shift[0]) {|x| abbreviate(x)}
+            component << " )"
           when :literal
-            string << %Q{ "#{path.shift}" }
+            component << quote_literal(parsed.shift)
           else
-            string << "/" unless string.size == 0
-            string << "UNKNOWN("
-            string << op.inspect
-            string << ")"
+            component << "UNKNOWN("
+            component << op.inspect
+            component << ")"
           end
         end
-        string = "/"+string if document
-        return string
+        case components
+        when [""]
+          "/"
+        when ["", ""]
+          "//"
+        else
+          components.join("/")
+        end
       end
 
-      def expand( path )
-        path = path.kind_of?(String) ? parse( path ) : path
-        string = ""
+      def expand(path_or_parsed)
+        if path_or_parsed.kind_of?(String)
+          parsed = parse(path_or_parsed)
+        else
+          parsed = path_or_parsed
+        end
+        path = ""
         document = false
-        while path.size > 0
-          op = path.shift
+        while parsed.size > 0
+          op = parsed.shift
           case op
           when :node
-            string << "node()"
+            path << "node()"
           when :attribute, :child, :following, :following_sibling,
                 :ancestor, :ancestor_or_self, :descendant, :descendant_or_self,
                 :namespace, :preceding, :preceding_sibling, :self, :parent
-            string << "/" unless string.size == 0
-            string << op.to_s.tr("_", "-")
-            string << "::"
+            path << "/" unless path.size == 0
+            path << op.to_s.tr("_", "-")
+            path << "::"
           when :any
-            string << "*"
+            path << "*"
           when :qname
-            prefix = path.shift
-            name = path.shift
-            string << prefix+":" if prefix.size > 0
-            string << name
+            prefix = parsed.shift
+            name = parsed.shift
+            path << prefix+":" if prefix.size > 0
+            path << name
           when :predicate
-            string << '['
-            string << predicate_to_string( path.shift ) { |x| expand(x) }
-            string << ']'
+            path << '['
+            path << predicate_to_path( parsed.shift ) { |x| expand(x) }
+            path << ']'
           when :document
             document = true
           else
-            string << "/" unless string.size == 0
-            string << "UNKNOWN("
-            string << op.inspect
-            string << ")"
+            path << "UNKNOWN("
+            path << op.inspect
+            path << ")"
           end
         end
-        string = "/"+string if document
-        return string
+        path = "/"+path if document
+        path
       end
 
-      def predicate_to_string( path, &block )
-        string = ""
-        case path[0]
+      def predicate_to_path(parsed, &block)
+        path = ""
+        case parsed[0]
         when :and, :or, :mult, :plus, :minus, :neq, :eq, :lt, :gt, :lteq, :gteq, :div, :mod, :union
-          op = path.shift
+          op = parsed.shift
           case op
           when :eq
             op = "="
@@ -150,36 +192,50 @@ module REXML
           when :union
             op = "|"
           end
-          left = predicate_to_string( path.shift, &block )
-          right = predicate_to_string( path.shift, &block )
-          string << " "
-          string << left
-          string << " "
-          string << op.to_s
-          string << " "
-          string << right
-          string << " "
+          left = predicate_to_path( parsed.shift, &block )
+          right = predicate_to_path( parsed.shift, &block )
+          path << left
+          path << " "
+          path << op.to_s
+          path << " "
+          path << right
         when :function
-          path.shift
-          name = path.shift
-          string << name
-          string << "( "
-          string << predicate_to_string( path.shift, &block )
-          string << " )"
+          parsed.shift
+          name = parsed.shift
+          path << name
+          path << "("
+          parsed.shift.each_with_index do |argument, i|
+            path << ", " if i > 0
+            path << predicate_to_path(argument, &block)
+          end
+          path << ")"
         when :literal
-          path.shift
-          string << " "
-          string << path.shift.inspect
-          string << " "
+          parsed.shift
+          path << quote_literal(parsed.shift)
         else
-          string << " "
-          string << yield( path )
-          string << " "
+          path << yield( parsed )
         end
-        return string.squeeze(" ")
+        return path.squeeze(" ")
       end
+      # For backward compatibility
+      alias_method :preciate_to_string, :predicate_to_path
 
       private
+      def quote_literal( literal )
+        case literal
+        when String
+          # XPath 1.0 does not support escape characters.
+          # Assumes literal does not contain both single and double quotes.
+          if literal.include?("'")
+            "\"#{literal}\""
+          else
+            "'#{literal}'"
+          end
+        else
+          literal.inspect
+        end
+      end
+
       #LocationPath
       #  | RelativeLocationPath
       #  | '/' RelativeLocationPath?
@@ -229,24 +285,28 @@ module REXML
               path = path[1..-1]
             end
           else
+            path_before_axis_specifier = path
+            parsed_not_abberviated = []
             if path[0] == ?@
-              parsed << :attribute
+              parsed_not_abberviated << :attribute
               path = path[1..-1]
               # Goto Nodetest
             elsif path =~ AXIS
-              parsed << $1.tr('-','_').intern
+              parsed_not_abberviated << $1.tr('-','_').intern
               path = $'
               # Goto Nodetest
             else
-              parsed << :child
+              parsed_not_abberviated << :child
             end
 
-            n = []
-            path = NodeTest( path, n)
-
-            path = Predicate( path, n )
+            path_before_node_test = path
+            path = NodeTest(path, parsed_not_abberviated)
+            if path == path_before_node_test
+              return path_before_axis_specifier
+            end
+            path = Predicate(path, parsed_not_abberviated)
 
-            parsed.concat(n)
+            parsed.concat(parsed_not_abberviated)
           end
 
           original_path = path
@@ -301,7 +361,9 @@ module REXML
         when PI
           path = $'
           literal = nil
-          if path !~ /^\s*\)/
+          if path =~ /^\s*\)/
+            path = $'
+          else
             path =~ LITERAL
             literal = $1
             path = $'
@@ -545,7 +607,9 @@ module REXML
       #| PrimaryExpr
       def FilterExpr path, parsed
         n = []
-        path = PrimaryExpr( path, n )
+        path_before_primary_expr = path
+        path = PrimaryExpr(path, n)
+        return path_before_primary_expr if path == path_before_primary_expr
         path = Predicate(path, n)
         parsed.concat(n)
         path

data/lib/rexml/rexml.rb

@@ -1,30 +1,37 @@
 # -*- coding: utf-8 -*-
 # frozen_string_literal: false
-# REXML is an XML toolkit for Ruby[http://www.ruby-lang.org], in Ruby.
-#
-# REXML is a _pure_ Ruby, XML 1.0 conforming,
-# non-validating[http://www.w3.org/TR/2004/REC-xml-20040204/#sec-conformance]
-# toolkit with an intuitive API.  REXML passes 100% of the non-validating Oasis
-# tests[http://www.oasis-open.org/committees/xml-conformance/xml-test-suite.shtml],
-# and provides tree, stream, SAX2, pull, and lightweight APIs.  REXML also
-# includes a full XPath[http://www.w3c.org/tr/xpath] 1.0 implementation. Since
-# Ruby 1.8, REXML is included in the standard Ruby distribution.
-#
-# Main page:: http://www.germane-software.com/software/rexml
-# Author:: Sean Russell <serATgermaneHYPHENsoftwareDOTcom>
-# Date:: 2008/019
-# Version:: 3.1.7.3
-#
-# This API documentation can be downloaded from the REXML home page, or can
-# be accessed online[http://www.germane-software.com/software/rexml_doc]
-#
-# A tutorial is available in the REXML distribution in docs/tutorial.html,
-# or can be accessed
-# online[http://www.germane-software.com/software/rexml/docs/tutorial.html]
+#
+# \Module \REXML provides classes and methods for parsing,
+# editing, and generating XML.
+#
+# == Implementation
+#
+# \REXML:
+# - Is pure Ruby.
+# - Provides tree, stream, SAX2, pull, and lightweight APIs.
+# - Conforms to {XML version 1.0}[https://www.w3.org/TR/REC-xml/].
+# - Fully implements {XPath version 1.0}[http://www.w3c.org/tr/xpath].
+# - Is {non-validating}[https://www.w3.org/TR/xml/].
+# - Passes 100% of the non-validating {Oasis tests}[http://www.oasis-open.org/committees/xml-conformance/xml-test-suite.shtml].
+#
+# == In a Hurry?
+#
+# If you're somewhat familiar with XML
+# and have a particular task in mind,
+# you may want to see {the tasks pages}[doc/rexml/tasks/tocs/master_toc_rdoc.html].
+#
+# == API
+#
+# Among the most important classes for using \REXML are:
+# - REXML::Document.
+# - REXML::Element.
+#
+# There's also an {REXML tutorial}[doc/rexml/tutorial_rdoc.html].
+#
 module REXML
   COPYRIGHT = "Copyright © 2001-2008 Sean Russell <ser@germane-software.com>"
   DATE = "2008/019"
-  VERSION = "3.2.3"
+  VERSION = "3.3.8"
   REVISION = ""
 
   Copyright = COPYRIGHT