require 'rexml/parseexception'
require 'rexml/source'

module REXML
	module Parsers
		# = Using the Pull Parser
		# <em>This API is experimental, and subject to change.</em>
		#  parser = PullParser.new( "<a>text<b att='val'/>txet</a>" )
		#  while parser.has_next?
		#    res = parser.next
		#    puts res[1]['att'] if res.start_tag? and res[0] == 'b'
		#  end
		# See the PullEvent class for information on the content of the results.
		# The data is identical to the arguments passed for the various events to
		# the StreamListener API.
		#
		# Notice that:
		#  parser = PullParser.new( "<a>BAD DOCUMENT" )
		#  while parser.has_next?
		#    res = parser.next
		#    raise res[1] if res.error?
		#  end
		#
		# Nat Price gave me some good ideas for the API.
		class BaseParser
			NCNAME_STR= '[\w:][\-\w\d.]*'
			NAME_STR= "(?:#{NCNAME_STR}:)?#{NCNAME_STR}"

			NAMECHAR = '[\-\w\d\.:]'
			NAME = "([\\w:]#{NAMECHAR}*)"
			NMTOKEN = "(?:#{NAMECHAR})+"
			NMTOKENS = "#{NMTOKEN}(\\s+#{NMTOKEN})*"
			REFERENCE = "(?:&#{NAME};|&#\\d+;|&#x[0-9a-fA-F]+;)"
			REFERENCE_RE = /#{REFERENCE}/

			DOCTYPE_START = /\A\s*<!DOCTYPE\s/um
			DOCTYPE_PATTERN = /\s*<!DOCTYPE\s+(.*?)(\[|>)/um
			ATTRIBUTE_PATTERN = /\s*(#{NAME_STR})\s*=\s*(["'])(.*?)\2/um
			COMMENT_START = /\A<!--/u
			COMMENT_PATTERN = /<!--(.*?)-->/um
			CDATA_START = /\A<!\[CDATA\[/u
			CDATA_END = /^\s*\]\s*>/um
			CDATA_PATTERN = /<!\[CDATA\[(.*?)\]\]>/um
			XMLDECL_START = /\A<\?xml\s/u;
			XMLDECL_PATTERN = /<\?xml\s+(.*?)\?>*/um
			INSTRUCTION_START = /\A<\?/u
			INSTRUCTION_PATTERN = /<\?(.*?)(\s+.*?)?\?>/um
			TAG_MATCH = /^<((?>#{NAME_STR}))\s*((?>\s+#{NAME_STR}\s*=\s*(["']).*?\3)*)\s*(\/)?>/um
			CLOSE_MATCH = /^\s*<\/(#{NAME_STR})\s*>/um

			VERSION = /\bversion\s*=\s*["'](.*?)['"]/um
			ENCODING = /\bencoding=["'](.*?)['"]/um
			STANDALONE = /\bstandalone=["'](.*?)['"]/um

			ENTITY_START = /^\s*<!ENTITY/
			IDENTITY = /^([!\*\w\-]+)(\s+#{NCNAME_STR})?(\s+["'].*?['"])?(\s+['"].*?["'])?/u
			ELEMENTDECL_START = /^\s*<!ELEMENT/um
			ELEMENTDECL_PATTERN = /^\s*(<!ELEMENT.*?)>/um
			ENUMERATION = "\\(\\s*#{NMTOKEN}(?:\\s*\\|\\s*#{NMTOKEN})*\\s*\\)"
			NOTATIONTYPE = "NOTATION\\s+\\(\\s*#{NAME}(?:\\s*\\|\\s*#{NAME})*\\s*\\)"
			ENUMERATEDTYPE = "(?:(?:#{NOTATIONTYPE})|(?:#{ENUMERATION}))"
			ATTTYPE = "(CDATA|ID|IDREF|IDREFS|ENTITY|ENTITIES|NMTOKEN|NMTOKENS|#{ENUMERATEDTYPE})"
			ATTVALUE = "(?:\"((?:[^<&\"]|#{REFERENCE})*)\")|(?:'((?:[^<&']|#{REFERENCE})*)')"
			DEFAULTDECL = "(#REQUIRED|#IMPLIED|(?:(#FIXED\\s+)?#{ATTVALUE}))"
			ATTDEF = "\\s+#{NAME}\\s+#{ATTTYPE}\\s+#{DEFAULTDECL}"
			ATTDEF_RE = /#{ATTDEF}/
			ATTLISTDECL_START = /^\s*<!ATTLIST/um
			ATTLISTDECL_PATTERN = /^\s*<!ATTLIST\s+#{NAME}(?:#{ATTDEF})*\s*>/um
			NOTATIONDECL_START = /^\s*<!NOTATION/um
			PUBLIC = /^\s*<!NOTATION\s+(\w[\-\w]*)\s+(PUBLIC)\s+((["']).*?\4)\s*>/um
			SYSTEM = /^\s*<!NOTATION\s+(\w[\-\w]*)\s+(SYSTEM)\s+((["']).*?\4)\s*>/um

			TEXT_PATTERN = /\A([^<]*)/um

			# Entity constants
			PUBIDCHAR = "\x20\x0D\x0Aa-zA-Z0-9\\-()+,./:=?;!*@$_%#"
			SYSTEMLITERAL = %Q{((?:"[^"]*")|(?:'[^']*'))}
			PUBIDLITERAL = %Q{("[#{PUBIDCHAR}']*"|'[#{PUBIDCHAR}]*')}
			EXTERNALID = "(?:(?:(SYSTEM)\\s+#{SYSTEMLITERAL})|(?:(PUBLIC)\\s+#{PUBIDLITERAL}\\s+#{SYSTEMLITERAL}))"
			NDATADECL = "\\s+NDATA\\s+#{NAME}"
			PEREFERENCE = "%#{NAME};"
			ENTITYVALUE = %Q{((?:"(?:[^%&"]|#{PEREFERENCE}|#{REFERENCE})*")|(?:'([^%&']|#{PEREFERENCE}|#{REFERENCE})*'))}
			PEDEF = "(?:#{ENTITYVALUE}|#{EXTERNALID})"
			ENTITYDEF = "(?:#{ENTITYVALUE}|(?:#{EXTERNALID}(#{NDATADECL})?))"
			PEDECL = "<!ENTITY\\s+(%)\\s+#{NAME}\\s+#{PEDEF}\\s*>"
			GEDECL = "<!ENTITY\\s+#{NAME}\\s+#{ENTITYDEF}\\s*>"
			ENTITYDECL = /\s*(?:#{GEDECL})|(?:#{PEDECL})/um

			EREFERENCE = /&(?!#{NAME};)/

			DEFAULT_ENTITIES = { 
				'gt' => [/&gt;/, '&gt;', '>'], 
				'lt' => [/&lt;/, '&lt;', '<'], 
				'quot' => [/&quot;/, '&quot;', '"'], 
				"apos" => [/&apos;/, "&apos;", "'"] 
			}

			def initialize( source )
				self.stream = source
			end

			def stream=( source )
				if source.kind_of? String
					@source = Source.new(source)
				elsif source.kind_of? IO
					@source = IOSource.new(source)
				elsif source.kind_of? Source
					@source = source
				else
					raise "#{source.type} is not a valid input stream.  It must be \n"+
					"either a String, IO, or Source."
				end
				@closed = nil
				@document_status = nil
				@tags = []
				@stack = []
				@entities = []
			end

			# Returns true if there are no more events
			def empty?
				!has_next?
			end

			# Returns true if there are more events.  Synonymous with !empty?
			def has_next?
				@source.read if @source.buffer.size==0 and !@source.empty?
				(!@source.empty? and @source.buffer.strip.size>0) or @stack.size>0 or @closed
			end

			# Push an event back on the head of the stream.  This method
			# has (theoretically) infinite depth.
			def unshift token
				@stack.unshift(token)
			end

			# Peek at the +depth+ event in the stack.  The first element on the stack
			# is at depth 0.  If +depth+ is -1, will parse to the end of the input
			# stream and return the last event, which is always :end_document.
			# Be aware that this causes the stream to be parsed up to the +depth+ 
			# event, so you can effectively pre-parse the entire document (pull the 
			# entire thing into memory) using this method.  
			def peek depth=0
				raise 'Illegal argument "#{depth}"' if depth < -1
				temp = []
				if depth == -1
					temp.push(pull()) until empty?
				else
					while @stack.size+temp.size < depth+1
						temp.push(pull())
					end
				end
				@stack += temp if temp.size > 0
				@stack[depth]
			end

			# Returns the next event.  This is a +PullEvent+ object.
			def pull
				return [ :end_document ] if empty?
				if @closed
					x, @closed = @closed, nil
					return [ :end_element, x ]
				end
				return @stack.shift if @stack.size > 0
				@source.read if @source.buffer.size==0
				if @document_status == nil
					@source.match( /^\s*/um, true )
					word = @source.match( /^\s*(<.*?)>/um )
					word = word[1] unless word.nil?
					case word
					when COMMENT_START
						return [ :comment, @source.match( COMMENT_PATTERN, true )[1] ]
					when XMLDECL_START
						results = @source.match( XMLDECL_PATTERN, true )[1]
						version = VERSION.match( results )
						version = version[1] unless version.nil?
						encoding = ENCODING.match(results)
						encoding = encoding[1] unless encoding.nil?
						@source.encoding = encoding
						standalone = STANDALONE.match(results)
						standalone = standalone[1] unless standalone.nil?
						return [ :xmldecl, version, encoding, standalone]
					when INSTRUCTION_START
						return [ :processing_instruction, *@source.match(INSTRUCTION_PATTERN, true)[1,2] ]
					when DOCTYPE_START
						md = @source.match( DOCTYPE_PATTERN, true )
						identity = md[1]
						close = md[2]
						identity =~ IDENTITY
						name = $1
						raise "DOCTYPE is missing a name" if name.nil?
						pub_sys = $2.nil? ? nil : $2.strip
						long_name = $3.nil? ? nil : $3.strip
						uri = $4.nil? ? nil : $4.strip
						args = [ :start_doctype, name, pub_sys, long_name, uri ]
						if close == ">"
							@document_status = :after_doctype
							@source.read if @source.buffer.size==0
							md = @source.match(/^\s*/um, true)
							@stack << [ :end_doctype ]
						else
							@document_status = :in_doctype
						end
						return args
					else
						@document_status = :after_doctype
						@source.read if @source.buffer.size==0
						md = @source.match(/\s*/um, true)
					end
				end
				if @document_status == :in_doctype
					md = @source.match(/\s*(.*?>)/um)
					case md[1]
					when ELEMENTDECL_START
						return [ :elementdecl, @source.match( ELEMENTDECL_PATTERN, true )[1] ]
					when ENTITY_START
						match = @source.match( ENTITYDECL, true ).to_a.compact
						match[0] = :entitydecl
						ref = false
						if match[1] == '%'
							ref = true
							match.delete_at 1
						end
						# Now we have to sort out what kind of entity reference this is
						if match[2] == 'SYSTEM'
							# External reference
							match[3] = match[3][1..-2] # PUBID
							match.delete_at(4) if match.size > 4 # Chop out NDATA decl
							# match is [ :entity, name, SYSTEM, pubid(, ndata)? ]
						elsif match[2] == 'PUBLIC'
							# External reference
							match[3] = match[3][1..-2] # PUBID
							match[4] = match[4][1..-2] # HREF
							# match is [ :entity, name, PUBLIC, pubid, href ]
						else
							match[2] = match[2][1..-2]
							match.pop if match.size == 4
							# match is [ :entity, name, value ]
						end
						match << '%' if ref
						return match
					when ATTLISTDECL_START
						md = @source.match( ATTLISTDECL_PATTERN, true )
						raise REXML::ParseException.new( "Bad ATTLIST declaration!", @source ) if md.nil?
						element = md[1]
						contents = md[0]

						pairs = {}
						values = md[0].scan( ATTDEF_RE )
						values.each do |attdef|
							unless attdef[3] == "#IMPLIED"
								attdef.compact!
								val = attdef[3]
								val = attdef[4] if val == "#FIXED "
								pairs[attdef[0]] = val
							end
						end
						return [ :attlistdecl, element, pairs, contents ]
					when NOTATIONDECL_START
						md = nil
						if @source.match( PUBLIC )
							md = @source.match( PUBLIC, true )
						elsif @source.match( SYSTEM )
							md = @source.match( SYSTEM, true )
						else
							raise REXML::ParseException.new( "error parsing notation: no matching pattern", @source )
						end
						return [ :notationdecl, md[1], md[2], md[3] ]
					when CDATA_END
						@document_status = :after_doctype
						@source.match( CDATA_END, true )
						return [ :end_doctype ]
					end
				end
				begin 
					if @source.buffer[0] == ?<
						if @source.buffer[1] == ?/
							last_tag = @tags.pop
							md = @source.match( CLOSE_MATCH, true )
							raise REXML::ParseException.new( "Missing end tag for '#{last_tag}' "+
								"(got \"#{md[1]}\")", @source) unless last_tag == md[1]
							return [ :end_element, last_tag ]
						elsif @source.buffer[1] == ?!
							md = @source.match(/\A(\s*[^>]*>)/um)
							#puts "SOURCE BUFFER = #{source.buffer}, #{source.buffer.size}"
							raise REXML::ParseException.new("Malformed node", @source) unless md
							case md[1]
							when CDATA_START
								return [ :cdata, @source.match( CDATA_PATTERN, true )[1] ]
							when COMMENT_START
								return [ :comment, @source.match( COMMENT_PATTERN, true )[1] ]
							else
								raise REXML::ParseException.new( "Declarations can only occur "+
								"in the doctype declaration.", @source)
							end
						elsif @source.buffer[1] == ??
							md = @source.match( INSTRUCTION_PATTERN, true )
							return [ :processing_instruction, md[1], md[2] ]
						else
							# Get the next tag
							md = @source.match(TAG_MATCH, true)
							raise REXML::ParseException.new("malformed XML: missing tag start", @source) unless md
							attrs = []
							if md[2].size > 0
								attrs = md[2].scan( ATTRIBUTE_PATTERN )
								raise REXML::ParseException.new( "error parsing attributes: [#{attrs.join ', '}], excess = \"#$'\"", @source) if $' and $'.strip.size > 0
							end
				
							if md[4]
								@closed = md[1]
							else
								@tags.push( md[1] )
							end
							attributes = {}
							attrs.each { |a,b,c| attributes[a] = c }
							return [ :start_element, md[1], attributes ]
						end
					else
						md = @source.match(TEXT_PATTERN, true)
						raise "no text to add" if md[0].length == 0
						# unnormalized = Text::unnormalize( md[1], self )
						# return PullEvent.new( :text, md[1], unnormalized )
						return [ :text, md[1] ]
					end
        rescue REXML::ParseException
          raise $!
				rescue Exception, NameError => error
					raise REXML::ParseException.new( "Exception parsing",
						@source, self, error )
				end
				return [ :dummy ]
			end

			def entity( reference, entities )
				value = nil
				value = entities[ reference ] if entities
				if not value
					value = DEFAULT_ENTITIES[ reference ]
					value = value[2] if value
				end
				unnormalize( value, entities ) if value
			end

			# Escapes all possible entities
			def normalize( input, entities=nil, entity_filter=nil )
				copy = input.clone
				# Doing it like this rather than in a loop improves the speed
				copy.gsub!( EREFERENCE, '&amp;' )
				entities.each do |key, value|
					copy.gsub!( value, "&#{key};" ) unless entity_filter and 
																			entity_filter.include?(entity)
				end if entities
				copy.gsub!( EREFERENCE, '&amp;' )
				DEFAULT_ENTITIES.each do |key, value|
					copy.gsub!( value[2], value[1] )
				end
				copy
			end

			# Unescapes all possible entities
			def unnormalize( string, entities=nil, filter=nil )
				rv = string.clone
				rv.gsub!( /\r\n?/, "\n" )
				matches = rv.scan( REFERENCE_RE )
				return rv if matches.size == 0
				rv.gsub!( /&#0*((?:\d+)|(?:x[a-fA-F0-9]+));/ ) {|m|
					m=$1
					m = "0#{m}" if m[0] == ?x
					[Integer(m)].pack('U*')
				}
				matches.collect!{|x|x[0]}.compact!
				if matches.size > 0
					matches.each do |entity_reference|
						unless filter and filter.include?(entity_reference)
							entity_value = entity( entity_reference, entities )
							if entity_value
								re = /&#{entity_reference};/
								rv.gsub!( re, entity_value )
							end
						end
					end
					matches.each do |entity_reference|
						unless filter and filter.include?(entity_reference)
							er = DEFAULT_ENTITIES[entity_reference]
							rv.gsub!( er[0], er[2] ) if er
						end
					end
					rv.gsub!( /&amp;/, '&' )
				end
				rv
			end
		end
	end
end