Chris@210: module CodeRay
Chris@210: module Scanners
Chris@210: 
Chris@210:   class CSS < Scanner
Chris@210: 
Chris@210:     register_for :css
Chris@210: 
Chris@210:     KINDS_NOT_LOC = [
Chris@210:       :comment,
Chris@210:       :class, :pseudo_class, :type,
Chris@210:       :constant, :directive,
Chris@210:       :key, :value, :operator, :color, :float,
Chris@210:       :error, :important,
Chris@210:     ]
Chris@210:     
Chris@210:     module RE
Chris@210:       Hex = /[0-9a-fA-F]/
Chris@210:       Unicode = /\\#{Hex}{1,6}(?:\r\n|\s)?/ # differs from standard because it allows uppercase hex too
Chris@210:       Escape = /#{Unicode}|\\[^\r\n\f0-9a-fA-F]/
Chris@210:       NMChar = /[-_a-zA-Z0-9]|#{Escape}/
Chris@210:       NMStart = /[_a-zA-Z]|#{Escape}/
Chris@210:       NL = /\r\n|\r|\n|\f/
Chris@210:       String1 = /"(?:[^\n\r\f\\"]|\\#{NL}|#{Escape})*"?/  # FIXME: buggy regexp
Chris@210:       String2 = /'(?:[^\n\r\f\\']|\\#{NL}|#{Escape})*'?/  # FIXME: buggy regexp
Chris@210:       String = /#{String1}|#{String2}/
Chris@210: 
Chris@210:       HexColor = /#(?:#{Hex}{6}|#{Hex}{3})/
Chris@210:       Color = /#{HexColor}/
Chris@210: 
Chris@210:       Num = /-?(?:[0-9]+|[0-9]*\.[0-9]+)/
Chris@210:       Name = /#{NMChar}+/
Chris@210:       Ident = /-?#{NMStart}#{NMChar}*/
Chris@210:       AtKeyword = /@#{Ident}/
Chris@210:       Percentage = /#{Num}%/
Chris@210: 
Chris@210:       reldimensions = %w[em ex px]
Chris@210:       absdimensions = %w[in cm mm pt pc]
Chris@210:       Unit = Regexp.union(*(reldimensions + absdimensions))
Chris@210: 
Chris@210:       Dimension = /#{Num}#{Unit}/
Chris@210: 
Chris@210:       Comment = %r! /\* (?: .*? \*/ | .* ) !mx
Chris@210:       Function = /(?:url|alpha)\((?:[^)\n\r\f]|\\\))*\)?/
Chris@210: 
Chris@210:       Id = /##{Name}/
Chris@210:       Class = /\.#{Name}/
Chris@210:       PseudoClass = /:#{Name}/
Chris@210:       AttributeSelector = /\[[^\]]*\]?/
Chris@210: 
Chris@210:     end
Chris@210: 
Chris@210:     def scan_tokens tokens, options
Chris@210:       
Chris@210:       value_expected = nil
Chris@210:       states = [:initial]
Chris@210: 
Chris@210:       until eos?
Chris@210: 
Chris@210:         kind = nil
Chris@210:         match = nil
Chris@210: 
Chris@210:         if scan(/\s+/)
Chris@210:           kind = :space
Chris@210: 
Chris@210:         elsif case states.last
Chris@210:           when :initial, :media
Chris@210:             if scan(/(?>#{RE::Ident})(?!\()|\*/ox)
Chris@210:               kind = :type
Chris@210:             elsif scan RE::Class
Chris@210:               kind = :class
Chris@210:             elsif scan RE::Id
Chris@210:               kind = :constant
Chris@210:             elsif scan RE::PseudoClass
Chris@210:               kind = :pseudo_class
Chris@210:             elsif match = scan(RE::AttributeSelector)
Chris@210:               # TODO: Improve highlighting inside of attribute selectors.
Chris@210:               tokens << [:open, :string]
Chris@210:               tokens << [match[0,1], :delimiter]
Chris@210:               tokens << [match[1..-2], :content] if match.size > 2
Chris@210:               tokens << [match[-1,1], :delimiter] if match[-1] == ?]
Chris@210:               tokens << [:close, :string]
Chris@210:               next
Chris@210:             elsif match = scan(/@media/)
Chris@210:               kind = :directive
Chris@210:               states.push :media_before_name
Chris@210:             end
Chris@210:           
Chris@210:           when :block
Chris@210:             if scan(/(?>#{RE::Ident})(?!\()/ox)
Chris@210:               if value_expected
Chris@210:                 kind = :value
Chris@210:               else
Chris@210:                 kind = :key
Chris@210:               end
Chris@210:             end
Chris@210: 
Chris@210:           when :media_before_name
Chris@210:             if scan RE::Ident
Chris@210:               kind = :type
Chris@210:               states[-1] = :media_after_name
Chris@210:             end
Chris@210:           
Chris@210:           when :media_after_name
Chris@210:             if scan(/\{/)
Chris@210:               kind = :operator
Chris@210:               states[-1] = :media
Chris@210:             end
Chris@210:           
Chris@210:           when :comment
Chris@210:             if scan(/(?:[^*\s]|\*(?!\/))+/)
Chris@210:               kind = :comment
Chris@210:             elsif scan(/\*\//)
Chris@210:               kind = :comment
Chris@210:               states.pop
Chris@210:             elsif scan(/\s+/)
Chris@210:               kind = :space
Chris@210:             end
Chris@210: 
Chris@210:           else
Chris@210:             raise_inspect 'Unknown state', tokens
Chris@210: 
Chris@210:           end
Chris@210: 
Chris@210:         elsif scan(/\/\*/)
Chris@210:           kind = :comment
Chris@210:           states.push :comment
Chris@210: 
Chris@210:         elsif scan(/\{/)
Chris@210:           value_expected = false
Chris@210:           kind = :operator
Chris@210:           states.push :block
Chris@210: 
Chris@210:         elsif scan(/\}/)
Chris@210:           value_expected = false
Chris@210:           if states.last == :block || states.last == :media
Chris@210:             kind = :operator
Chris@210:             states.pop
Chris@210:           else
Chris@210:             kind = :error
Chris@210:           end
Chris@210: 
Chris@210:         elsif match = scan(/#{RE::String}/o)
Chris@210:           tokens << [:open, :string]
Chris@210:           tokens << [match[0, 1], :delimiter]
Chris@210:           tokens << [match[1..-2], :content] if match.size > 2
Chris@210:           tokens << [match[-1, 1], :delimiter] if match.size >= 2
Chris@210:           tokens << [:close, :string]
Chris@210:           next
Chris@210: 
Chris@210:         elsif match = scan(/#{RE::Function}/o)
Chris@210:           tokens << [:open, :string]
Chris@210:           start = match[/^\w+\(/]
Chris@210:           tokens << [start, :delimiter]
Chris@210:           if match[-1] == ?)
Chris@210:             tokens << [match[start.size..-2], :content]
Chris@210:             tokens << [')', :delimiter]
Chris@210:           else
Chris@210:             tokens << [match[start.size..-1], :content]
Chris@210:           end
Chris@210:           tokens << [:close, :string]
Chris@210:           next
Chris@210: 
Chris@210:         elsif scan(/(?: #{RE::Dimension} | #{RE::Percentage} | #{RE::Num} )/ox)
Chris@210:           kind = :float
Chris@210: 
Chris@210:         elsif scan(/#{RE::Color}/o)
Chris@210:           kind = :color
Chris@210: 
Chris@210:         elsif scan(/! *important/)
Chris@210:           kind = :important
Chris@210: 
Chris@210:         elsif scan(/rgb\([^()\n]*\)?/)
Chris@210:           kind = :color
Chris@210: 
Chris@210:         elsif scan(/#{RE::AtKeyword}/o)
Chris@210:           kind = :directive
Chris@210: 
Chris@210:         elsif match = scan(/ [+>:;,.=()\/] /x)
Chris@210:           if match == ':'
Chris@210:             value_expected = true
Chris@210:           elsif match == ';'
Chris@210:             value_expected = false
Chris@210:           end
Chris@210:           kind = :operator
Chris@210: 
Chris@210:         else
Chris@210:           getch
Chris@210:           kind = :error
Chris@210: 
Chris@210:         end
Chris@210: 
Chris@210:         match ||= matched
Chris@210:         if $CODERAY_DEBUG and not kind
Chris@210:           raise_inspect 'Error token %p in line %d' %
Chris@210:             [[match, kind], line], tokens
Chris@210:         end
Chris@210:         raise_inspect 'Empty token', tokens unless match
Chris@210: 
Chris@210:         tokens << [match, kind]
Chris@210: 
Chris@210:       end
Chris@210: 
Chris@210:       tokens
Chris@210:     end
Chris@210: 
Chris@210:   end
Chris@210: 
Chris@210: end
Chris@210: end