diff options
Diffstat (limited to 'lib')
-rw-r--r-- | lib/rexml/doctype.rb | 71 | ||||
-rw-r--r-- | lib/rexml/parsers/baseparser.rb | 178 | ||||
-rw-r--r-- | lib/rexml/rexml.rb | 2 |
3 files changed, 192 insertions, 59 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 |