Source

RubyLearning / ProjectTrak / vendor / rails / actionpack / lib / action_controller / vendor / html-scanner / html / selector.rb

Full commit
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
#--
# Copyright (c) 2006 Assaf Arkin (http://labnotes.org)
# Under MIT and/or CC By license.
#++

module HTML

  # Selects HTML elements using CSS 2 selectors.
  #
  # The +Selector+ class uses CSS selector expressions to match and select
  # HTML elements.
  #
  # For example:
  #   selector = HTML::Selector.new "form.login[action=/login]"
  # creates a new selector that matches any +form+ element with the class
  # +login+ and an attribute +action+ with the value <tt>/login</tt>.
  #
  # === Matching Elements
  #
  # Use the #match method to determine if an element matches the selector.
  #
  # For simple selectors, the method returns an array with that element,
  # or +nil+ if the element does not match. For complex selectors (see below)
  # the method returns an array with all matched elements, of +nil+ if no
  # match found.
  #
  # For example:
  #   if selector.match(element)
  #     puts "Element is a login form"
  #   end
  #
  # === Selecting Elements
  #
  # Use the #select method to select all matching elements starting with
  # one element and going through all children in depth-first order.
  #
  # This method returns an array of all matching elements, an empty array
  # if no match is found
  #
  # For example:
  #   selector = HTML::Selector.new "input[type=text]"
  #   matches = selector.select(element)
  #   matches.each do |match|
  #     puts "Found text field with name #{match.attributes['name']}"
  #   end
  #
  # === Expressions
  #
  # Selectors can match elements using any of the following criteria:
  # * <tt>name</tt> -- Match an element based on its name (tag name).
  #   For example, <tt>p</tt> to match a paragraph. You can use <tt>*</tt>
  #   to match any element.
  # * <tt>#</tt><tt>id</tt> -- Match an element based on its identifier (the
  #   <tt>id</tt> attribute). For example, <tt>#</tt><tt>page</tt>.
  # * <tt>.class</tt> -- Match an element based on its class name, all
  #   class names if more than one specified.
  # * <tt>[attr]</tt> -- Match an element that has the specified attribute.
  # * <tt>[attr=value]</tt> -- Match an element that has the specified
  #   attribute and value. (More operators are supported see below)
  # * <tt>:pseudo-class</tt> -- Match an element based on a pseudo class,
  #   such as <tt>:nth-child</tt> and <tt>:empty</tt>.
  # * <tt>:not(expr)</tt> -- Match an element that does not match the
  #   negation expression.
  #
  # When using a combination of the above, the element name comes first
  # followed by identifier, class names, attributes, pseudo classes and
  # negation in any order. Do not separate these parts with spaces!
  # Space separation is used for descendant selectors.
  #
  # For example:
  #   selector = HTML::Selector.new "form.login[action=/login]"
  # The matched element must be of type +form+ and have the class +login+.
  # It may have other classes, but the class +login+ is required to match.
  # It must also have an attribute called +action+ with the value
  # <tt>/login</tt>.
  #
  # This selector will match the following element:
  #   <form class="login form" method="post" action="/login">
  # but will not match the element:
  #   <form method="post" action="/logout">
  #
  # === Attribute Values
  #
  # Several operators are supported for matching attributes:
  # * <tt>name</tt> -- The element must have an attribute with that name.
  # * <tt>name=value</tt> -- The element must have an attribute with that
  #   name and value.
  # * <tt>name^=value</tt> -- The attribute value must start with the
  #   specified value.
  # * <tt>name$=value</tt> -- The attribute value must end with the
  #   specified value.
  # * <tt>name*=value</tt> -- The attribute value must contain the
  #   specified value.
  # * <tt>name~=word</tt> -- The attribute value must contain the specified
  #   word (space separated).
  # * <tt>name|=word</tt> -- The attribute value must start with specified
  #   word.
  #
  # For example, the following two selectors match the same element:
  #   #my_id
  #   [id=my_id]
  # and so do the following two selectors:
  #   .my_class
  #   [class~=my_class]
  #
  # === Alternatives, siblings, children
  #
  # Complex selectors use a combination of expressions to match elements:
  # * <tt>expr1 expr2</tt> -- Match any element against the second expression
  #   if it has some parent element that matches the first expression.
  # * <tt>expr1 > expr2</tt> -- Match any element against the second expression
  #   if it is the child of an element that matches the first expression.
  # * <tt>expr1 + expr2</tt> -- Match any element against the second expression
  #   if it immediately follows an element that matches the first expression.
  # * <tt>expr1 ~ expr2</tt> -- Match any element against the second expression
  #   that comes after an element that matches the first expression.
  # * <tt>expr1, expr2</tt> -- Match any element against the first expression,
  #   or against the second expression.
  #
  # Since children and sibling selectors may match more than one element given
  # the first element, the #match method may return more than one match.
  #
  # === Pseudo classes
  #
  # Pseudo classes were introduced in CSS 3. They are most often used to select
  # elements in a given position:
  # * <tt>:root</tt> -- Match the element only if it is the root element
  #   (no parent element).
  # * <tt>:empty</tt> -- Match the element only if it has no child elements,
  #   and no text content.
  # * <tt>:only-child</tt> -- Match the element if it is the only child (element)
  #   of its parent element.
  # * <tt>:only-of-type</tt> -- Match the element if it is the only child (element)
  #   of its parent element and its type.
  # * <tt>:first-child</tt> -- Match the element if it is the first child (element)
  #   of its parent element.
  # * <tt>:first-of-type</tt> -- Match the element if it is the first child (element)
  #   of its parent element of its type.
  # * <tt>:last-child</tt> -- Match the element if it is the last child (element)
  #   of its parent element.
  # * <tt>:last-of-type</tt> -- Match the element if it is the last child (element)
  #   of its parent element of its type.
  # * <tt>:nth-child(b)</tt> -- Match the element if it is the b-th child (element)
  #   of its parent element. The value <tt>b</tt> specifies its index, starting with 1.
  # * <tt>:nth-child(an+b)</tt> -- Match the element if it is the b-th child (element)
  #   in each group of <tt>a</tt> child elements of its parent element.
  # * <tt>:nth-child(-an+b)</tt> -- Match the element if it is the first child (element)
  #   in each group of <tt>a</tt> child elements, up to the first <tt>b</tt> child
  #   elements of its parent element.
  # * <tt>:nth-child(odd)</tt> -- Match element in the odd position (i.e. first, third).
  #   Same as <tt>:nth-child(2n+1)</tt>.
  # * <tt>:nth-child(even)</tt> -- Match element in the even position (i.e. second,
  #   fourth). Same as <tt>:nth-child(2n+2)</tt>.
  # * <tt>:nth-of-type(..)</tt> -- As above, but only counts elements of its type.
  # * <tt>:nth-last-child(..)</tt> -- As above, but counts from the last child.
  # * <tt>:nth-last-of-type(..)</tt> -- As above, but counts from the last child and
  #   only elements of its type.
  # * <tt>:not(selector)</tt> -- Match the element only if the element does not
  #   match the simple selector.
  #
  # As you can see, <tt>:nth-child<tt> pseudo class and its variant can get quite
  # tricky and the CSS specification doesn't do a much better job explaining it.
  # But after reading the examples and trying a few combinations, it's easy to
  # figure out.
  #
  # For example:
  #   table tr:nth-child(odd)
  # Selects every second row in the table starting with the first one.
  #
  #   div p:nth-child(4)
  # Selects the fourth paragraph in the +div+, but not if the +div+ contains
  # other elements, since those are also counted.
  #
  #   div p:nth-of-type(4)
  # Selects the fourth paragraph in the +div+, counting only paragraphs, and
  # ignoring all other elements.
  #
  #   div p:nth-of-type(-n+4)
  # Selects the first four paragraphs, ignoring all others.
  #
  # And you can always select an element that matches one set of rules but
  # not another using <tt>:not</tt>. For example:
  #   p:not(.post)
  # Matches all paragraphs that do not have the class <tt>.post</tt>.
  #   
  # === Substitution Values
  #
  # You can use substitution with identifiers, class names and element values.
  # A substitution takes the form of a question mark (<tt>?</tt>) and uses the
  # next value in the argument list following the CSS expression.
  #
  # The substitution value may be a string or a regular expression. All other
  # values are converted to strings.
  #
  # For example:
  #   selector = HTML::Selector.new "#?", /^\d+$/
  # matches any element whose identifier consists of one or more digits.
  #
  # See http://www.w3.org/TR/css3-selectors/
  class Selector


    # An invalid selector.
    class InvalidSelectorError < StandardError #:nodoc:
    end


    class << self

      # :call-seq:
      #   Selector.for_class(cls) => selector
      #
      # Creates a new selector for the given class name.
      def for_class(cls)
        self.new([".?", cls])
      end


      # :call-seq:
      #   Selector.for_id(id) => selector
      #
      # Creates a new selector for the given id.
      def for_id(id)
        self.new(["#?", id])
      end

    end


    # :call-seq:
    #   Selector.new(string, [values ...]) => selector
    #
    # Creates a new selector from a CSS 2 selector expression.
    #
    # The first argument is the selector expression. All other arguments
    # are used for value substitution.
    #
    # Throws InvalidSelectorError is the selector expression is invalid.
    def initialize(selector, *values)
      raise ArgumentError, "CSS expression cannot be empty" if selector.empty?
      @source = ""
      values = values[0] if values.size == 1 && values[0].is_a?(Array)

      # We need a copy to determine if we failed to parse, and also
      # preserve the original pass by-ref statement.
      statement = selector.strip.dup

      # Create a simple selector, along with negation.
      simple_selector(statement, values).each { |name, value| instance_variable_set("@#{name}", value) }

      @alternates = []
      @depends = nil

      # Alternative selector.
      if statement.sub!(/^\s*,\s*/, "")
        second = Selector.new(statement, values)
        @alternates << second
        # If there are alternate selectors, we group them in the top selector.
        if alternates = second.instance_variable_get(:@alternates)
          second.instance_variable_set(:@alternates, [])
          @alternates.concat alternates
        end
        @source << " , " << second.to_s
      # Sibling selector: create a dependency into second selector that will
      # match element immediately following this one.
      elsif statement.sub!(/^\s*\+\s*/, "")
        second = next_selector(statement, values)
        @depends = lambda do |element, first|
          if element = next_element(element)
            second.match(element, first)
          end
        end
        @source << " + " << second.to_s
      # Adjacent selector: create a dependency into second selector that will
      # match all elements following this one.
      elsif statement.sub!(/^\s*~\s*/, "")
        second = next_selector(statement, values)
        @depends = lambda do |element, first|
          matches = []
          while element = next_element(element)
            if subset = second.match(element, first)
              if first && !subset.empty?
                matches << subset.first
                break
              else
                matches.concat subset
              end
            end
          end
          matches.empty? ? nil : matches
        end
        @source << " ~ " << second.to_s
      # Child selector: create a dependency into second selector that will
      # match a child element of this one.
      elsif statement.sub!(/^\s*>\s*/, "")
        second = next_selector(statement, values)
        @depends = lambda do |element, first|
          matches = []
          element.children.each do |child|
            if child.tag? && subset = second.match(child, first)
              if first && !subset.empty?
                matches << subset.first
                break
              else
                matches.concat subset
              end
            end
          end
          matches.empty? ? nil : matches
        end
        @source << " > " << second.to_s
      # Descendant selector: create a dependency into second selector that
      # will match all descendant elements of this one. Note,
      elsif statement =~ /^\s+\S+/ && statement != selector
        second = next_selector(statement, values)
        @depends = lambda do |element, first|
          matches = []
          stack = element.children.reverse
          while node = stack.pop
            next unless node.tag?
            if subset = second.match(node, first)
              if first && !subset.empty?
                matches << subset.first
                break
              else
                matches.concat subset
              end
            elsif children = node.children
              stack.concat children.reverse
            end
          end
          matches.empty? ? nil : matches
        end
        @source << " " << second.to_s
      else
        # The last selector is where we check that we parsed
        # all the parts.
        unless statement.empty? || statement.strip.empty?
          raise ArgumentError, "Invalid selector: #{statement}"
        end
      end
    end


    # :call-seq:
    #   match(element, first?) => array or nil
    #
    # Matches an element against the selector.
    #
    # For a simple selector this method returns an array with the
    # element if the element matches, nil otherwise.
    #
    # For a complex selector (sibling and descendant) this method
    # returns an array with all matching elements, nil if no match is
    # found.
    #
    # Use +first_only=true+ if you are only interested in the first element.
    #
    # For example:
    #   if selector.match(element)
    #     puts "Element is a login form"
    #   end
    def match(element, first_only = false)
      # Match element if no element name or element name same as element name
      if matched = (!@tag_name || @tag_name == element.name)
        # No match if one of the attribute matches failed
        for attr in @attributes
          if element.attributes[attr[0]] !~ attr[1]
            matched = false
            break
          end
        end
      end

      # Pseudo class matches (nth-child, empty, etc).
      if matched
        for pseudo in @pseudo
          unless pseudo.call(element)
            matched = false
            break
          end
        end
      end

      # Negation. Same rules as above, but we fail if a match is made.
      if matched && @negation
        for negation in @negation
          if negation[:tag_name] == element.name
            matched = false
          else
            for attr in negation[:attributes]
              if element.attributes[attr[0]] =~ attr[1]
                matched = false
                break
              end
            end
          end
          if matched
            for pseudo in negation[:pseudo]
              if pseudo.call(element)
                matched = false
                break
              end
            end
          end
          break unless matched
        end
      end

      # If element matched but depends on another element (child,
      # sibling, etc), apply the dependent matches instead.
      if matched && @depends
        matches = @depends.call(element, first_only)
      else
        matches = matched ? [element] : nil
      end

      # If this selector is part of the group, try all the alternative
      # selectors (unless first_only).
      if !first_only || !matches
        @alternates.each do |alternate|
          break if matches && first_only
          if subset = alternate.match(element, first_only)
            if matches
              matches.concat subset
            else
              matches = subset
            end
          end
        end
      end

      matches
    end


    # :call-seq:
    #   select(root) => array
    #
    # Selects and returns an array with all matching elements, beginning
    # with one node and traversing through all children depth-first.
    # Returns an empty array if no match is found.
    #
    # The root node may be any element in the document, or the document
    # itself.
    #
    # For example:
    #   selector = HTML::Selector.new "input[type=text]"
    #   matches = selector.select(element)
    #   matches.each do |match|
    #     puts "Found text field with name #{match.attributes['name']}"
    #   end
    def select(root)
      matches = []
      stack = [root]
      while node = stack.pop
        if node.tag? && subset = match(node, false)
          subset.each do |match|
            matches << match unless matches.any? { |item| item.equal?(match) }
          end
        elsif children = node.children
          stack.concat children.reverse
        end
      end
      matches
    end


    # Similar to #select but returns the first matching element. Returns +nil+
    # if no element matches the selector.
    def select_first(root)
      stack = [root]
      while node = stack.pop
        if node.tag? && subset = match(node, true)
          return subset.first if !subset.empty?
        elsif children = node.children
          stack.concat children.reverse
        end
      end
      nil
    end


    def to_s #:nodoc:
      @source
    end


    # Return the next element after this one. Skips sibling text nodes.
    #
    # With the +name+ argument, returns the next element with that name,
    # skipping other sibling elements.
    def next_element(element, name = nil)
      if siblings = element.parent.children
        found = false
        siblings.each do |node|
          if node.equal?(element)
            found = true
          elsif found && node.tag?
            return node if (name.nil? || node.name == name)
          end
        end
      end
      nil
    end


  protected


    # Creates a simple selector given the statement and array of
    # substitution values.
    #
    # Returns a hash with the values +tag_name+, +attributes+,
    # +pseudo+ (classes) and +negation+.
    #
    # Called the first time with +can_negate+ true to allow
    # negation. Called a second time with false since negation
    # cannot be negated.
    def simple_selector(statement, values, can_negate = true)
      tag_name = nil
      attributes = []
      pseudo = []
      negation = []

      # Element name. (Note that in negation, this can come at
      # any order, but for simplicity we allow if only first).
      statement.sub!(/^(\*|[[:alpha:]][\w\-]*)/) do |match|
        match.strip!
        tag_name = match.downcase unless match == "*"
        @source << match
        "" # Remove
      end

      # Get identifier, class, attribute name, pseudo or negation.
      while true
        # Element identifier.
        next if statement.sub!(/^#(\?|[\w\-]+)/) do |match|
          id = $1
          if id == "?"
            id = values.shift
          end
          @source << "##{id}"
          id = Regexp.new("^#{Regexp.escape(id.to_s)}$") unless id.is_a?(Regexp)
          attributes << ["id", id]
          "" # Remove
        end

        # Class name.
        next if statement.sub!(/^\.([\w\-]+)/) do |match|
          class_name = $1
          @source << ".#{class_name}"
          class_name = Regexp.new("(^|\s)#{Regexp.escape(class_name)}($|\s)") unless class_name.is_a?(Regexp)
          attributes << ["class", class_name]
          "" # Remove
        end

        # Attribute value.
        next if statement.sub!(/^\[\s*([[:alpha:]][\w\-]*)\s*((?:[~|^$*])?=)?\s*('[^']*'|"[^*]"|[^\]]*)\s*\]/) do |match|
          name, equality, value = $1, $2, $3
          if value == "?"
            value = values.shift
          else
            # Handle single and double quotes.
            value.strip!
            if (value[0] == ?" || value[0] == ?') && value[0] == value[-1]
              value = value[1..-2]
            end
          end
          @source << "[#{name}#{equality}'#{value}']"
          attributes << [name.downcase.strip, attribute_match(equality, value)]
          "" # Remove
        end

        # Root element only.
        next if statement.sub!(/^:root/) do |match|
          pseudo << lambda do |element|
            element.parent.nil? || !element.parent.tag?
          end
          @source << ":root"
          "" # Remove
        end

        # Nth-child including last and of-type.
        next if statement.sub!(/^:nth-(last-)?(child|of-type)\((odd|even|(\d+|\?)|(-?\d*|\?)?n([+\-]\d+|\?)?)\)/) do |match|
          reverse = $1 == "last-"
          of_type = $2 == "of-type"
          @source << ":nth-#{$1}#{$2}("
          case $3
            when "odd"
              pseudo << nth_child(2, 1, of_type, reverse)
              @source << "odd)"
            when "even"
              pseudo << nth_child(2, 2, of_type, reverse)
              @source << "even)"
            when /^(\d+|\?)$/  # b only
              b = ($1 == "?" ? values.shift : $1).to_i
              pseudo << nth_child(0, b, of_type, reverse)
              @source << "#{b})"
            when /^(-?\d*|\?)?n([+\-]\d+|\?)?$/
              a = ($1 == "?" ? values.shift :
                   $1 == "" ? 1 : $1 == "-" ? -1 : $1).to_i
              b = ($2 == "?" ? values.shift : $2).to_i
              pseudo << nth_child(a, b, of_type, reverse)
              @source << (b >= 0 ? "#{a}n+#{b})" : "#{a}n#{b})")
            else
              raise ArgumentError, "Invalid nth-child #{match}"
          end
          "" # Remove
        end
        # First/last child (of type).
        next if statement.sub!(/^:(first|last)-(child|of-type)/) do |match|
          reverse = $1 == "last"
          of_type = $2 == "of-type"
          pseudo << nth_child(0, 1, of_type, reverse)
          @source << ":#{$1}-#{$2}"
          "" # Remove
        end
        # Only child (of type).
        next if statement.sub!(/^:only-(child|of-type)/) do |match|
          of_type = $1 == "of-type"
          pseudo << only_child(of_type)
          @source << ":only-#{$1}"
          "" # Remove
        end

        # Empty: no child elements or meaningful content (whitespaces
        # are ignored).
        next if statement.sub!(/^:empty/) do |match|
          pseudo << lambda do |element|
            empty = true
            for child in element.children
              if child.tag? || !child.content.strip.empty?
                empty = false
                break
              end
            end
            empty
          end
          @source << ":empty"
          "" # Remove
        end
        # Content: match the text content of the element, stripping
        # leading and trailing spaces.
        next if statement.sub!(/^:content\(\s*(\?|'[^']*'|"[^"]*"|[^)]*)\s*\)/) do |match|
          content = $1
          if content == "?"
            content = values.shift
          elsif (content[0] == ?" || content[0] == ?') && content[0] == content[-1]
            content = content[1..-2]
          end
          @source << ":content('#{content}')"
          content = Regexp.new("^#{Regexp.escape(content.to_s)}$") unless content.is_a?(Regexp)
          pseudo << lambda do |element|
            text = ""
            for child in element.children
              unless child.tag?
                text << child.content
              end
            end
            text.strip =~ content
          end
          "" # Remove
        end

        # Negation. Create another simple selector to handle it.
        if statement.sub!(/^:not\(\s*/, "")
          raise ArgumentError, "Double negatives are not missing feature" unless can_negate
          @source << ":not("
          negation << simple_selector(statement, values, false)
          raise ArgumentError, "Negation not closed" unless statement.sub!(/^\s*\)/, "")
          @source << ")"
          next
        end

        # No match: moving on.
        break
      end

      # Return hash. The keys are mapped to instance variables.
      {:tag_name=>tag_name, :attributes=>attributes, :pseudo=>pseudo, :negation=>negation}
    end


    # Create a regular expression to match an attribute value based
    # on the equality operator (=, ^=, |=, etc).
    def attribute_match(equality, value)
      regexp = value.is_a?(Regexp) ? value : Regexp.escape(value.to_s)
      case equality
        when "=" then
          # Match the attribute value in full
          Regexp.new("^#{regexp}$")
        when "~=" then
          # Match a space-separated word within the attribute value
          Regexp.new("(^|\s)#{regexp}($|\s)")
        when "^="
          # Match the beginning of the attribute value
          Regexp.new("^#{regexp}")
        when "$="
          # Match the end of the attribute value
          Regexp.new("#{regexp}$")
        when "*="
          # Match substring of the attribute value
          regexp.is_a?(Regexp) ? regexp : Regexp.new(regexp)
        when "|=" then
          # Match the first space-separated item of the attribute value
          Regexp.new("^#{regexp}($|\s)")
        else
          raise InvalidSelectorError, "Invalid operation/value" unless value.empty?
          # Match all attributes values (existence check)
          //
      end
    end


    # Returns a lambda that can match an element against the nth-child
    # pseudo class, given the following arguments:
    # * +a+ -- Value of a part.
    # * +b+ -- Value of b part.
    # * +of_type+ -- True to test only elements of this type (of-type).
    # * +reverse+ -- True to count in reverse order (last-).
    def nth_child(a, b, of_type, reverse)
      # a = 0 means select at index b, if b = 0 nothing selected
      return lambda { |element| false } if a == 0 && b == 0
      # a < 0 and b < 0 will never match against an index
      return lambda { |element| false } if a < 0 && b < 0
      b = a + b + 1 if b < 0   # b < 0 just picks last element from each group
      b -= 1 unless b == 0  # b == 0 is same as b == 1, otherwise zero based
      lambda do |element|
        # Element must be inside parent element.
        return false unless element.parent && element.parent.tag?
        index = 0
        # Get siblings, reverse if counting from last.
        siblings = element.parent.children
        siblings = siblings.reverse if reverse
        # Match element name if of-type, otherwise ignore name.
        name = of_type ? element.name : nil
        found = false
        for child in siblings
          # Skip text nodes/comments.
          if child.tag? && (name == nil || child.name == name)
            if a == 0
              # Shortcut when a == 0 no need to go past count
              if index == b
                found = child.equal?(element)
                break
              end
            elsif a < 0
              # Only look for first b elements
              break if index > b
              if child.equal?(element)
                found = (index % a) == 0
                break
              end
            else
              # Otherwise, break if child found and count ==  an+b
              if child.equal?(element)
                found = (index % a) == b
                break
              end
            end
            index += 1
          end
        end
        found
      end
    end


    # Creates a only child lambda. Pass +of-type+ to only look at
    # elements of its type.
    def only_child(of_type)
      lambda do |element|
        # Element must be inside parent element.
        return false unless element.parent && element.parent.tag?
        name = of_type ? element.name : nil
        other = false
        for child in element.parent.children
          # Skip text nodes/comments.
          if child.tag? && (name == nil || child.name == name)
            unless child.equal?(element)
              other = true
              break
            end
          end
        end
        !other
      end
    end


    # Called to create a dependent selector (sibling, descendant, etc).
    # Passes the remainder of the statement that will be reduced to zero
    # eventually, and array of substitution values.
    #
    # This method is called from four places, so it helps to put it here
    # for reuse. The only logic deals with the need to detect comma
    # separators (alternate) and apply them to the selector group of the
    # top selector.
    def next_selector(statement, values)
      second = Selector.new(statement, values)
      # If there are alternate selectors, we group them in the top selector.
      if alternates = second.instance_variable_get(:@alternates)
        second.instance_variable_set(:@alternates, [])
        @alternates.concat alternates
      end
      second
    end

  end


  # See HTML::Selector.new
  def self.selector(statement, *values)
    Selector.new(statement, *values)
  end


  class Tag

    def select(selector, *values)
      selector = HTML::Selector.new(selector, values)
      selector.select(self)
    end

  end

end