regexp_parser 1.7.0 → 2.8.1

data/lib/regexp_parser/expression/classes/{set → character_set}/range.rb

@@ -1,23 +1,19 @@
 module Regexp::Expression
   class CharacterSet < Regexp::Expression::Subexpression
     class Range < Regexp::Expression::Subexpression
-      def starts_at
-        expressions.first.starts_at
+      def ts
+        (head = expressions.first) ? head.ts : @ts
       end
-      alias :ts :starts_at
 
       def <<(exp)
-        complete? && raise("Can't add more than 2 expressions to a Range")
+        complete? and raise Regexp::Parser::Error,
+          "Can't add more than 2 expressions to a Range"
         super
       end
 
       def complete?
         count == 2
       end
-
-      def to_s(_format = :full)
-        expressions.join(text)
-      end
     end
   end
 end

data/lib/regexp_parser/expression/classes/{set.rb → character_set.rb}

@@ -19,9 +19,8 @@ module Regexp::Expression
     def close
       self.closed = true
     end
-
-    def to_s(format = :full)
-      "#{text}#{'^' if negated?}#{expressions.join}]#{quantifier_affix(format)}"
-    end
   end
+
+  # alias for symmetry between token symbol and Expression class name
+  Set = CharacterSet
 end # module Regexp::Expression

data/lib/regexp_parser/expression/classes/{type.rb → character_type.rb}

@@ -1,5 +1,4 @@
 module Regexp::Expression
-
   module CharacterType
     class Base < Regexp::Expression::Base; end
 
@@ -15,5 +14,4 @@ module Regexp::Expression
     class Linebreak        < CharacterType::Base; end
     class ExtendedGrapheme < CharacterType::Base; end
   end
-
 end

data/lib/regexp_parser/expression/classes/conditional.rb

@@ -1,6 +1,6 @@
 module Regexp::Expression
   module Conditional
-    class TooManyBranches < StandardError
+    class TooManyBranches < Regexp::Parser::Error
       def initialize
         super('The conditional expression has more than 2 branches')
       end
@@ -15,6 +15,11 @@ module Regexp::Expression
         ref = text.tr("'<>()", "")
         ref =~ /\D/ ? ref : Integer(ref)
       end
+
+      def initialize_copy(orig)
+        self.referenced_expression = orig.referenced_expression.dup
+        super
+      end
     end
 
     class Branch < Regexp::Expression::Sequence; end
@@ -26,9 +31,9 @@ module Regexp::Expression
         expressions.last << exp
       end
 
-      def add_sequence(active_opts = {})
+      def add_sequence(active_opts = {}, params = { ts: 0 })
         raise TooManyBranches.new if branches.length == 2
-        params = { conditional_level: conditional_level + 1 }
+        params = params.merge({ conditional_level: conditional_level + 1 })
         Branch.add_to(self, params, active_opts)
       end
       alias :branch :add_sequence
@@ -50,8 +55,9 @@ module Regexp::Expression
         condition.reference
       end
 
-      def to_s(format = :full)
-        "#{text}#{condition}#{branches.join('|')})#{quantifier_affix(format)}"
+      def initialize_copy(orig)
+        self.referenced_expression = orig.referenced_expression.dup
+        super
       end
     end
   end

data/lib/regexp_parser/expression/classes/{escape.rb → escape_sequence.rb}

@@ -1,16 +1,21 @@
 module Regexp::Expression
   module EscapeSequence
     class Base < Regexp::Expression::Base
-      require 'yaml'
-
-      def char
-        # poor man's unescape without using eval
-        YAML.load(%Q(---\n"#{text}"\n))
-      end
-
       def codepoint
         char.ord
       end
+
+      if ''.respond_to?(:undump)
+        def char
+          %("#{text}").undump
+        end
+      else
+        # poor man's unescape without using eval
+        require 'yaml'
+        def char
+          YAML.load(%Q(---\n"#{text}"\n))
+        end
+      end
     end
 
     class Literal < EscapeSequence::Base
@@ -91,4 +96,7 @@ module Regexp::Expression
       end
     end
   end
+
+  # alias for symmetry between Token::* and Expression::*
+  Escape = EscapeSequence
 end

data/lib/regexp_parser/expression/classes/free_space.rb

@@ -1,17 +1,17 @@
 module Regexp::Expression
-
   class FreeSpace < Regexp::Expression::Base
-    def quantify(token, text, min = nil, max = nil, mode = :greedy)
-      raise "Can not quantify a free space object"
+    def quantify(*_args)
+      raise Regexp::Parser::Error, 'Can not quantify a free space object'
     end
   end
 
-  class Comment < Regexp::Expression::FreeSpace; end
+  class Comment < Regexp::Expression::FreeSpace
+  end
 
   class WhiteSpace < Regexp::Expression::FreeSpace
     def merge(exp)
+      warn("#{self.class}##{__method__} is deprecated and will be removed in v3.0.0.")
       text << exp.text
     end
   end
-
 end

data/lib/regexp_parser/expression/classes/group.rb

@@ -1,27 +1,45 @@
 module Regexp::Expression
   module Group
     class Base < Regexp::Expression::Subexpression
-      def to_s(format = :full)
-        "#{text}#{expressions.join})#{quantifier_affix(format)}"
-      end
+    end
 
-      def capturing?; false end
+    class Passive < Group::Base
+      attr_writer :implicit
+
+      def initialize(*)
+        @implicit = false
+        super
+      end
 
-      def comment?; false end
+      def implicit?
+        @implicit
+      end
     end
 
-    class Atomic  < Group::Base; end
-    class Passive < Group::Base; end
     class Absence < Group::Base; end
+    class Atomic  < Group::Base; end
+    # TODO: should split off OptionsSwitch in v3.0.0. Maybe even make it no
+    # longer inherit from Group because it is effectively a terminal expression.
     class Options < Group::Base
       attr_accessor :option_changes
+
+      def initialize_copy(orig)
+        self.option_changes = orig.option_changes.dup
+        super
+      end
+
+      def quantify(*args)
+        if token == :options_switch
+          raise Regexp::Parser::Error, 'Can not quantify an option switch'
+        else
+          super
+        end
+      end
     end
 
     class Capture < Group::Base
       attr_accessor :number, :number_at_level
       alias identifier number
-
-      def capturing?; true end
     end
 
     class Named < Group::Capture
@@ -33,18 +51,13 @@ module Regexp::Expression
         super
       end
 
-      def initialize_clone(orig)
+      def initialize_copy(orig)
         @name = orig.name.dup
         super
       end
     end
 
     class Comment < Group::Base
-      def to_s(_format = :full)
-        text.dup
-      end
-
-      def comment?; true end
     end
   end
 

data/lib/regexp_parser/expression/classes/keep.rb

@@ -1,5 +1,7 @@
 module Regexp::Expression
   module Keep
+    # TOOD: in regexp_parser v3.0.0 this should possibly be a Subexpression
+    #       that contains all expressions to its left.
     class Mark < Regexp::Expression::Base; end
   end
 end

data/lib/regexp_parser/expression/classes/literal.rb

@@ -1,7 +1,3 @@
 module Regexp::Expression
-
-  class Literal < Regexp::Expression::Base
-    # Obviously nothing special here, yet.
-  end
-
+  class Literal < Regexp::Expression::Base; end
 end

data/lib/regexp_parser/expression/classes/posix_class.rb

@@ -5,7 +5,11 @@ module Regexp::Expression
     end
 
     def name
-      token.to_s
+      text[/\w+/]
     end
   end
+
+  # alias for symmetry between token symbol and Expression class name
+  Posixclass    = PosixClass
+  Nonposixclass = PosixClass
 end

data/lib/regexp_parser/expression/classes/root.rb

@@ -1,24 +1,9 @@
 module Regexp::Expression
-
   class Root < Regexp::Expression::Subexpression
-    # TODO: this override is here for backwards compatibility, remove in 2.0.0
-    def initialize(*args)
-      unless args.first.is_a?(Regexp::Token)
-        warn('WARNING: Root.new without a Token argument is deprecated and '\
-             'will be removed in 2.0.0. Use Root.build for the old behavior.')
-        return super(self.class.build_token, *args)
-      end
-      super
-    end
-
-    class << self
-      def build(options = {})
-        new(build_token, options)
-      end
-
-      def build_token
-        Regexp::Token.new(:expression, :root, '', 0)
-      end
+    def self.build(options = {})
+      warn "`#{self.class}.build(options)` is deprecated and will raise in "\
+           "regexp_parser v3.0.0. Please use `.construct(options: options)`."
+      construct(options: options)
     end
   end
 end

data/lib/regexp_parser/expression/classes/{property.rb → unicode_property.rb}

@@ -1,5 +1,4 @@
 module Regexp::Expression
-
   module UnicodeProperty
     class Base < Regexp::Expression::Base
       def negative?
@@ -7,11 +6,11 @@ module Regexp::Expression
       end
 
       def name
-        text =~ /\A\\[pP]\{([^}]+)\}\z/; $1
+        text[/\A\\[pP]\{([^}]+)\}\z/, 1]
       end
 
       def shortcut
-        (Regexp::Scanner.short_prop_map.rassoc(token.to_s) || []).first
+        Regexp::Scanner.short_prop_map.key(token.to_s)
       end
     end
 
@@ -117,4 +116,7 @@ module Regexp::Expression
     class Block   < UnicodeProperty::Base; end
   end
 
+  # alias for symmetry between token symbol and Expression class name
+  Property    = UnicodeProperty
+  Nonproperty = UnicodeProperty
 end # module Regexp::Expression

data/lib/regexp_parser/expression/methods/construct.rb

@@ -0,0 +1,41 @@
+module Regexp::Expression
+  module Shared
+    module ClassMethods
+      # Convenience method to init a valid Expression without a Regexp::Token
+      def construct(params = {})
+        attrs = construct_defaults.merge(params)
+        options = attrs.delete(:options)
+        token_args = Regexp::TOKEN_KEYS.map { |k| attrs.delete(k) }
+        token = Regexp::Token.new(*token_args)
+        raise ArgumentError, "unsupported attribute(s): #{attrs}" if attrs.any?
+
+        new(token, options)
+      end
+
+      def construct_defaults
+        if self == Root
+          { type: :expression, token: :root, ts: 0 }
+        elsif self < Sequence
+          { type: :expression, token: :sequence }
+        else
+          { type: token_class::Type }
+        end.merge(level: 0, set_level: 0, conditional_level: 0, text: '')
+      end
+
+      def token_class
+        if self == Root || self < Sequence
+          nil # no token class because these objects are Parser-generated
+        # TODO: synch exp class, token class & type names for this in v3.0.0
+        elsif self == CharacterType::Any
+          Regexp::Syntax::Token::Meta
+        else
+          Regexp::Syntax::Token.const_get(name.split('::')[2])
+        end
+      end
+    end
+
+    def token_class
+      self.class.token_class
+    end
+  end
+end

data/lib/regexp_parser/expression/methods/human_name.rb

@@ -0,0 +1,43 @@
+module Regexp::Expression
+  module Shared
+    # default implementation, e.g. "atomic group", "hex escape", "word type", ..
+    def human_name
+      [token, type].compact.join(' ').tr('_', ' ')
+    end
+  end
+
+  Alternation.class_eval                       { def human_name; 'alternation'                 end }
+  Alternative.class_eval                       { def human_name; 'alternative'                 end }
+  Anchor::BOL.class_eval                       { def human_name; 'beginning of line'           end }
+  Anchor::BOS.class_eval                       { def human_name; 'beginning of string'         end }
+  Anchor::EOL.class_eval                       { def human_name; 'end of line'                 end }
+  Anchor::EOS.class_eval                       { def human_name; 'end of string'               end }
+  Anchor::EOSobEOL.class_eval                  { def human_name; 'newline-ready end of string' end }
+  Anchor::MatchStart.class_eval                { def human_name; 'match start'                 end }
+  Anchor::NonWordBoundary.class_eval           { def human_name; 'no word boundary'            end }
+  Anchor::WordBoundary.class_eval              { def human_name; 'word boundary'               end }
+  Assertion::Lookahead.class_eval              { def human_name; 'lookahead'                   end }
+  Assertion::Lookbehind.class_eval             { def human_name; 'lookbehind'                  end }
+  Assertion::NegativeLookahead.class_eval      { def human_name; 'negative lookahead'          end }
+  Assertion::NegativeLookbehind.class_eval     { def human_name; 'negative lookbehind'         end }
+  Backreference::Name.class_eval               { def human_name; 'backreference by name'       end }
+  Backreference::NameCall.class_eval           { def human_name; 'subexpression call by name'  end }
+  Backreference::Number.class_eval             { def human_name; 'backreference'               end }
+  Backreference::NumberRelative.class_eval     { def human_name; 'relative backreference'      end }
+  Backreference::NumberCall.class_eval         { def human_name; 'subexpression call'          end }
+  Backreference::NumberCallRelative.class_eval { def human_name; 'relative subexpression call' end }
+  CharacterSet::IntersectedSequence.class_eval { def human_name; 'intersected sequence'        end }
+  CharacterSet::Intersection.class_eval        { def human_name; 'intersection'                end }
+  CharacterSet::Range.class_eval               { def human_name; 'character range'             end }
+  CharacterType::Any.class_eval                { def human_name; 'match-all'                   end }
+  Comment.class_eval                           { def human_name; 'comment'                     end }
+  Conditional::Branch.class_eval               { def human_name; 'conditional branch'          end }
+  Conditional::Condition.class_eval            { def human_name; 'condition'                   end }
+  Conditional::Expression.class_eval           { def human_name; 'conditional'                 end }
+  Group::Capture.class_eval                    { def human_name; "capture group #{number}"     end }
+  Group::Named.class_eval                      { def human_name; 'named capture group'         end }
+  Keep::Mark.class_eval                        { def human_name; 'keep-mark lookbehind'        end }
+  Literal.class_eval                           { def human_name; 'literal'                     end }
+  Root.class_eval                              { def human_name; 'root'                        end }
+  WhiteSpace.class_eval                        { def human_name; 'free space'                  end }
+end

data/lib/regexp_parser/expression/methods/match_length.rb

@@ -10,7 +10,7 @@ class Regexp::MatchLength
     self.exp_class = exp.class
     self.min_rep = exp.repetitions.min
     self.max_rep = exp.repetitions.max
-    if base = opts[:base]
+    if (base = opts[:base])
       self.base_min = base
       self.base_max = base
       self.reify = ->{ '.' * base }
@@ -32,7 +32,7 @@ class Regexp::MatchLength
     end
   end
 
-  def endless_each(&block)
+  def endless_each
     return enum_for(__method__) unless block_given?
     (min..max).each { |num| yield(num) if include?(num) }
   end
@@ -63,16 +63,20 @@ class Regexp::MatchLength
   end
 
   def to_re
-    "(?:#{reify.call}){#{min_rep},#{max_rep unless max_rep == Float::INFINITY}}"
+    /(?:#{reify.call}){#{min_rep},#{max_rep unless max_rep == Float::INFINITY}}/
   end
 
   private
 
   attr_accessor :base_min, :base_max, :min_rep, :max_rep, :exp_class, :reify
 
-  def test_regexp
-    @test_regexp ||= Regexp.new("^#{to_re}$").tap do |regexp|
-      regexp.respond_to?(:match?) || def regexp.match?(str); !!match(str) end
+  if Regexp.method_defined?(:match?) # ruby >= 2.4
+    def test_regexp
+      @test_regexp ||= /^#{to_re}$/
+    end
+  else
+    def test_regexp
+      @test_regexp ||= /^#{to_re}$/.tap { |r| def r.match?(s); !!match(s) end }
     end
   end
 end
@@ -112,7 +116,7 @@ module Regexp::Expression
     end
 
     def inner_match_length
-      dummy = Regexp::Expression::Root.build
+      dummy = Regexp::Expression::Root.construct
       dummy.expressions = expressions.map(&:clone)
       dummy.quantifier = quantifier && quantifier.clone
       dummy.match_length

data/lib/regexp_parser/expression/methods/parts.rb

@@ -0,0 +1,23 @@
+module Regexp::Expression
+  module Shared
+    # default implementation
+    def parts
+      [text.dup]
+    end
+
+    private
+
+    def intersperse(expressions, separator)
+      expressions.flat_map { |exp| [exp, separator] }.slice(0...-1)
+    end
+  end
+
+  CharacterSet.class_eval            { def parts; ["#{text}#{'^' if negated?}", *expressions, ']']        end }
+  CharacterSet::Range.class_eval     { def parts; intersperse(expressions, text.dup)                      end }
+  Conditional::Expression.class_eval { def parts; [text.dup, condition, *intersperse(branches, '|'), ')'] end }
+  Group::Base.class_eval             { def parts; [text.dup, *expressions, ')']                           end }
+  Group::Passive.class_eval          { def parts; implicit? ? expressions : super                         end }
+  Group::Comment.class_eval          { def parts; [text.dup]                                              end }
+  Subexpression.class_eval           { def parts; expressions                                             end }
+  SequenceOperation.class_eval       { def parts; intersperse(expressions, text.dup)                      end }
+end

data/lib/regexp_parser/expression/methods/printing.rb

@@ -0,0 +1,26 @@
+module Regexp::Expression
+  module Shared
+    def inspect
+      [
+        "#<#{self.class}",
+        pretty_print_instance_variables.map { |v| " #{v}=#{instance_variable_get(v).inspect}" },
+        ">"
+      ].join
+    end
+
+    # Make pretty-print work despite #inspect implementation.
+    def pretty_print(q)
+      q.pp_object(self)
+    end
+
+    # Called by pretty_print (ruby/pp) and #inspect.
+    def pretty_print_instance_variables
+      [
+        (:@text unless text.to_s.empty?),
+        (:@quantifier if quantified?),
+        (:@options unless options.empty?),
+        (:@expressions unless terminal?),
+      ].compact
+    end
+  end
+end

data/lib/regexp_parser/expression/methods/strfregexp.rb

@@ -43,7 +43,7 @@ module Regexp::Expression
 
       # Order is important! Fields that use other fields in their
       # definition must appear before the fields they use.
-      part_keys = %w{a m b o i l x s e S y k c q Q z Z t ~t T >}
+      part_keys = %w[a m b o i l x s e S y k c q Q z Z t ~t T >]
       part.keys.each {|k| part[k] = "<?#{k}?>"}
 
       part['>'] = print_level ? ('  ' * (print_level + indent_offset)) : ''

data/lib/regexp_parser/expression/methods/tests.rb

@@ -1,5 +1,5 @@
 module Regexp::Expression
-  class Base
+  module Shared
 
     # Test if this expression has the given test_type, which can be either
     # a symbol or an array of symbols to check against the expression's type.
@@ -93,5 +93,51 @@ module Regexp::Expression
               "Array, Hash, or Symbol expected, #{scope.class.name} given"
       end
     end
+
+    # Deep-compare two expressions for equality.
+    #
+    # When changing the conditions, please make sure to update
+    # #pretty_print_instance_variables so that it includes all relevant values.
+    def ==(other)
+      self.class   == other.class &&
+        text       == other.text &&
+        quantifier == other.quantifier &&
+        options    == other.options &&
+        (terminal? || expressions == other.expressions)
+    end
+    alias :=== :==
+    alias :eql? :==
+
+    def optional?
+      quantified? && quantifier.min == 0
+    end
+
+    def quantified?
+      !quantifier.nil?
+    end
   end
+
+  Shared.class_eval                     { def terminal?; self.class.terminal? end }
+  Shared::ClassMethods.class_eval       { def terminal?; true  end }
+  Subexpression.instance_eval           { def terminal?; false end }
+
+  Shared.class_eval                     { def capturing?; self.class.capturing? end }
+  Shared::ClassMethods.class_eval       { def capturing?; false end }
+  Group::Capture.instance_eval          { def capturing?; true  end }
+
+  Shared.class_eval                     { def comment?; self.class.comment? end }
+  Shared::ClassMethods.class_eval       { def comment?; false end }
+  Comment.instance_eval                 { def comment?; true  end }
+  Group::Comment.instance_eval          { def comment?; true  end }
+
+  Shared.class_eval                     { def decorative?; self.class.decorative? end }
+  Shared::ClassMethods.class_eval       { def decorative?; false end }
+  FreeSpace.instance_eval               { def decorative?; true  end }
+  Group::Comment.instance_eval          { def decorative?; true  end }
+
+  Shared.class_eval                     { def referential?; self.class.referential? end }
+  Shared::ClassMethods.class_eval       { def referential?; false end }
+  Backreference::Base.instance_eval     { def referential?; true  end }
+  Conditional::Condition.instance_eval  { def referential?; true  end }
+  Conditional::Expression.instance_eval { def referential?; true  end }
 end

data/lib/regexp_parser/expression/methods/traverse.rb

@@ -1,6 +1,22 @@
 module Regexp::Expression
   class Subexpression < Regexp::Expression::Base
 
+    # Traverses the expression, passing each recursive child to the
+    # given block.
+    # If the block takes two arguments, the indices of the children within
+    # their parents are also passed to it.
+    def each_expression(include_self = false, &block)
+      return enum_for(__method__, include_self) unless block
+
+      if block.arity == 1
+        block.call(self) if include_self
+        each_expression_without_index(&block)
+      else
+        block.call(self, 0) if include_self
+        each_expression_with_index(&block)
+      end
+    end
+
     # Traverses the subexpression (depth-first, pre-order) and calls the given
     # block for each expression with three arguments; the traversal event,
     # the expression, and the index of the expression within its parent.
@@ -34,31 +50,31 @@ module Regexp::Expression
     end
     alias :walk :traverse
 
-    # Iterates over the expressions of this expression as an array, passing
-    # the expression and its index within its parent to the given block.
-    def each_expression(include_self = false, &block)
-      return enum_for(__method__, include_self) unless block_given?
-
-      traverse(include_self) do |event, exp, index|
-        yield(exp, index) unless event == :exit
-      end
-    end
-
     # Returns a new array with the results of calling the given block once
     # for every expression. If a block is not given, returns an array with
     # each expression and its level index as an array.
     def flat_map(include_self = false, &block)
-      result = []
+      case block && block.arity
+      when nil then each_expression(include_self).to_a
+      when 2   then each_expression(include_self).map(&block)
+      else          each_expression(include_self).map { |exp| block.call(exp) }
+      end
+    end
 
-      each_expression(include_self) do |exp, index|
-        if block_given?
-          result << yield(exp, index)
-        else
-          result << [exp, index]
-        end
+    protected
+
+    def each_expression_with_index(&block)
+      each_with_index do |exp, index|
+        block.call(exp, index)
+        exp.each_expression_with_index(&block) unless exp.terminal?
       end
+    end
 
-      result
+    def each_expression_without_index(&block)
+      each do |exp|
+        block.call(exp)
+        exp.each_expression_without_index(&block) unless exp.terminal?
+      end
     end
   end
 end