regexp_parser 2.1.1 → 2.9.2

data/lib/regexp_parser/parser.rb

@@ -1,5 +1,5 @@
-require 'regexp_parser/error'
-require 'regexp_parser/expression'
+require_relative 'error'
+require_relative 'expression'
 
 class Regexp::Parser
   include Regexp::Expression
@@ -18,12 +18,12 @@ class Regexp::Parser
     end
   end
 
-  def self.parse(input, syntax = "ruby/#{RUBY_VERSION}", options: nil, &block)
+  def self.parse(input, syntax = nil, options: nil, &block)
     new.parse(input, syntax, options: options, &block)
   end
 
-  def parse(input, syntax = "ruby/#{RUBY_VERSION}", options: nil, &block)
-    root = Root.build(extract_options(input, options))
+  def parse(input, syntax = nil, options: nil, &block)
+    root = Root.construct(options: extract_options(input, options))
 
     self.root = root
     self.node = root
@@ -35,10 +35,13 @@ class Regexp::Parser
 
     self.captured_group_counts = Hash.new(0)
 
-    Regexp::Lexer.scan(input, syntax, options: options) do |token|
+    Regexp::Lexer.scan(input, syntax, options: options, collect_tokens: false) do |token|
       parse_token(token)
     end
 
+    # Trigger recursive setting of #nesting_level, which reflects how deep
+    # a node is in the tree. Do this at the end to account for tree rewrites.
+    root.nesting_level = 0
     assign_referenced_expressions
 
     if block_given?
@@ -197,11 +200,11 @@ class Regexp::Parser
   end
 
   def captured_group_count_at_level
-    captured_group_counts[node.level]
+    captured_group_counts[node]
   end
 
   def count_captured_group
-    captured_group_counts[node.level] += 1
+    captured_group_counts[node] += 1
   end
 
   def close_group
@@ -229,10 +232,18 @@ class Regexp::Parser
       node << Backreference::NameRecursionLevel.new(token, active_opts)
     when :name_call
       node << Backreference::NameCall.new(token, active_opts)
-    when :number, :number_ref
+    when :number, :number_ref # TODO: split in v3.0.0
       node << Backreference::Number.new(token, active_opts)
     when :number_recursion_ref
-      node << Backreference::NumberRecursionLevel.new(token, active_opts)
+      node << Backreference::NumberRecursionLevel.new(token, active_opts).tap do |exp|
+        # TODO: should split off new token number_recursion_rel_ref and new
+        # class NumberRelativeRecursionLevel in v3.0.0 to get rid of this
+        if exp.text =~ /[<'][+-]/
+          assign_effective_number(exp)
+        else
+          exp.effective_number = exp.number
+        end
+      end
     when :number_call
       node << Backreference::NumberCall.new(token, active_opts)
     when :number_rel_ref
@@ -251,6 +262,8 @@ class Regexp::Parser
   def assign_effective_number(exp)
     exp.effective_number =
       exp.number + total_captured_group_count + (exp.number < 0 ? 1 : 0)
+    exp.effective_number > 0 ||
+      raise(ParserError, "Invalid reference: #{exp.reference}")
   end
 
   def conditional(token)
@@ -259,9 +272,9 @@ class Regexp::Parser
       nest_conditional(Conditional::Expression.new(token, active_opts))
     when :condition
       conditional_nesting.last.condition = Conditional::Condition.new(token, active_opts)
-      conditional_nesting.last.add_sequence(active_opts)
+      conditional_nesting.last.add_sequence(active_opts, { ts: token.te })
     when :separator
-      conditional_nesting.last.add_sequence(active_opts)
+      conditional_nesting.last.add_sequence(active_opts, { ts: token.te })
       self.node = conditional_nesting.last.branches.last
     when :close
       conditional_nesting.pop
@@ -286,17 +299,9 @@ class Regexp::Parser
   def nest(exp)
     nesting.push(exp)
     node << exp
-    update_transplanted_subtree(exp, node)
     self.node = exp
   end
 
-  # subtrees are transplanted to build Alternations, Intersections, Ranges
-  def update_transplanted_subtree(exp, new_parent)
-    exp.nesting_level = new_parent.nesting_level + 1
-    exp.respond_to?(:each) &&
-      exp.each { |subexp| update_transplanted_subtree(subexp, exp) }
-  end
-
   def escape(token)
     case token.token
 
@@ -317,6 +322,7 @@ class Regexp::Parser
 
     when :control
       if token.text =~ /\A(?:\\C-\\M|\\c\\M)/
+        # TODO: emit :meta_control_sequence token in v3.0.0
         node << EscapeSequence::MetaControl.new(token, active_opts)
       else
         node << EscapeSequence::Control.new(token, active_opts)
@@ -324,6 +330,7 @@ class Regexp::Parser
 
     when :meta_sequence
       if token.text =~ /\A\\M-\\[Cc]/
+        # TODO: emit :meta_control_sequence token in v3.0.0:
         node << EscapeSequence::MetaControl.new(token, active_opts)
       else
         node << EscapeSequence::Meta.new(token, active_opts)
@@ -344,11 +351,7 @@ class Regexp::Parser
     when :comment
       node << Comment.new(token, active_opts)
     when :whitespace
-      if node.last.is_a?(WhiteSpace)
-        node.last.merge(WhiteSpace.new(token, active_opts))
-      else
-        node << WhiteSpace.new(token, active_opts)
-      end
+      node << WhiteSpace.new(token, active_opts)
     else
       raise UnknownTokenError.new('FreeSpace', token)
     end
@@ -374,98 +377,99 @@ class Regexp::Parser
   end
 
   def sequence_operation(klass, token)
-    unless node.is_a?(klass)
+    unless node.instance_of?(klass)
       operator = klass.new(token, active_opts)
-      sequence = operator.add_sequence(active_opts)
+      sequence = operator.add_sequence(active_opts, { ts: token.ts })
       sequence.expressions = node.expressions
       node.expressions = []
       nest(operator)
     end
-    node.add_sequence(active_opts)
+    node.add_sequence(active_opts, { ts: token.te })
   end
 
   def posixclass(token)
     node << PosixClass.new(token, active_opts)
   end
 
-  include Regexp::Expression::UnicodeProperty
-  UPTokens = Regexp::Syntax::Token::UnicodeProperty
+  UP = Regexp::Expression::Property
+  UPTokens = Regexp::Syntax::Token::Property
 
   def property(token)
     case token.token
-    when :alnum;                  node << Alnum.new(token, active_opts)
-    when :alpha;                  node << Alpha.new(token, active_opts)
-    when :ascii;                  node << Ascii.new(token, active_opts)
-    when :blank;                  node << Blank.new(token, active_opts)
-    when :cntrl;                  node << Cntrl.new(token, active_opts)
-    when :digit;                  node << Digit.new(token, active_opts)
-    when :graph;                  node << Graph.new(token, active_opts)
-    when :lower;                  node << Lower.new(token, active_opts)
-    when :print;                  node << Print.new(token, active_opts)
-    when :punct;                  node << Punct.new(token, active_opts)
-    when :space;                  node << Space.new(token, active_opts)
-    when :upper;                  node << Upper.new(token, active_opts)
-    when :word;                   node << Word.new(token, active_opts)
-    when :xdigit;                 node << Xdigit.new(token, active_opts)
-    when :xposixpunct;            node << XPosixPunct.new(token, active_opts)
+    when :alnum;                  node << UP::Alnum.new(token, active_opts)
+    when :alpha;                  node << UP::Alpha.new(token, active_opts)
+    when :ascii;                  node << UP::Ascii.new(token, active_opts)
+    when :blank;                  node << UP::Blank.new(token, active_opts)
+    when :cntrl;                  node << UP::Cntrl.new(token, active_opts)
+    when :digit;                  node << UP::Digit.new(token, active_opts)
+    when :graph;                  node << UP::Graph.new(token, active_opts)
+    when :lower;                  node << UP::Lower.new(token, active_opts)
+    when :print;                  node << UP::Print.new(token, active_opts)
+    when :punct;                  node << UP::Punct.new(token, active_opts)
+    when :space;                  node << UP::Space.new(token, active_opts)
+    when :upper;                  node << UP::Upper.new(token, active_opts)
+    when :word;                   node << UP::Word.new(token, active_opts)
+    when :xdigit;                 node << UP::Xdigit.new(token, active_opts)
+    when :xposixpunct;            node << UP::XPosixPunct.new(token, active_opts)
 
     # only in Oniguruma (old rubies)
-    when :newline;                node << Newline.new(token, active_opts)
-
-    when :any;                    node << Any.new(token, active_opts)
-    when :assigned;               node << Assigned.new(token, active_opts)
-
-    when :letter;                 node << Letter::Any.new(token, active_opts)
-    when :cased_letter;           node << Letter::Cased.new(token, active_opts)
-    when :uppercase_letter;       node << Letter::Uppercase.new(token, active_opts)
-    when :lowercase_letter;       node << Letter::Lowercase.new(token, active_opts)
-    when :titlecase_letter;       node << Letter::Titlecase.new(token, active_opts)
-    when :modifier_letter;        node << Letter::Modifier.new(token, active_opts)
-    when :other_letter;           node << Letter::Other.new(token, active_opts)
-
-    when :mark;                   node << Mark::Any.new(token, active_opts)
-    when :combining_mark;         node << Mark::Combining.new(token, active_opts)
-    when :nonspacing_mark;        node << Mark::Nonspacing.new(token, active_opts)
-    when :spacing_mark;           node << Mark::Spacing.new(token, active_opts)
-    when :enclosing_mark;         node << Mark::Enclosing.new(token, active_opts)
-
-    when :number;                 node << Number::Any.new(token, active_opts)
-    when :decimal_number;         node << Number::Decimal.new(token, active_opts)
-    when :letter_number;          node << Number::Letter.new(token, active_opts)
-    when :other_number;           node << Number::Other.new(token, active_opts)
-
-    when :punctuation;            node << Punctuation::Any.new(token, active_opts)
-    when :connector_punctuation;  node << Punctuation::Connector.new(token, active_opts)
-    when :dash_punctuation;       node << Punctuation::Dash.new(token, active_opts)
-    when :open_punctuation;       node << Punctuation::Open.new(token, active_opts)
-    when :close_punctuation;      node << Punctuation::Close.new(token, active_opts)
-    when :initial_punctuation;    node << Punctuation::Initial.new(token, active_opts)
-    when :final_punctuation;      node << Punctuation::Final.new(token, active_opts)
-    when :other_punctuation;      node << Punctuation::Other.new(token, active_opts)
-
-    when :separator;              node << Separator::Any.new(token, active_opts)
-    when :space_separator;        node << Separator::Space.new(token, active_opts)
-    when :line_separator;         node << Separator::Line.new(token, active_opts)
-    when :paragraph_separator;    node << Separator::Paragraph.new(token, active_opts)
-
-    when :symbol;                 node << Symbol::Any.new(token, active_opts)
-    when :math_symbol;            node << Symbol::Math.new(token, active_opts)
-    when :currency_symbol;        node << Symbol::Currency.new(token, active_opts)
-    when :modifier_symbol;        node << Symbol::Modifier.new(token, active_opts)
-    when :other_symbol;           node << Symbol::Other.new(token, active_opts)
-
-    when :other;                  node << Codepoint::Any.new(token, active_opts)
-    when :control;                node << Codepoint::Control.new(token, active_opts)
-    when :format;                 node << Codepoint::Format.new(token, active_opts)
-    when :surrogate;              node << Codepoint::Surrogate.new(token, active_opts)
-    when :private_use;            node << Codepoint::PrivateUse.new(token, active_opts)
-    when :unassigned;             node << Codepoint::Unassigned.new(token, active_opts)
-
-    when *UPTokens::Age;          node << Age.new(token, active_opts)
-    when *UPTokens::Derived;      node << Derived.new(token, active_opts)
-    when *UPTokens::Emoji;        node << Emoji.new(token, active_opts)
-    when *UPTokens::Script;       node << Script.new(token, active_opts)
-    when *UPTokens::UnicodeBlock; node << Block.new(token, active_opts)
+    when :newline;                node << UP::Newline.new(token, active_opts)
+
+    when :any;                    node << UP::Any.new(token, active_opts)
+    when :assigned;               node << UP::Assigned.new(token, active_opts)
+
+    when :letter;                 node << UP::Letter::Any.new(token, active_opts)
+    when :cased_letter;           node << UP::Letter::Cased.new(token, active_opts)
+    when :uppercase_letter;       node << UP::Letter::Uppercase.new(token, active_opts)
+    when :lowercase_letter;       node << UP::Letter::Lowercase.new(token, active_opts)
+    when :titlecase_letter;       node << UP::Letter::Titlecase.new(token, active_opts)
+    when :modifier_letter;        node << UP::Letter::Modifier.new(token, active_opts)
+    when :other_letter;           node << UP::Letter::Other.new(token, active_opts)
+
+    when :mark;                   node << UP::Mark::Any.new(token, active_opts)
+    when :combining_mark;         node << UP::Mark::Combining.new(token, active_opts)
+    when :nonspacing_mark;        node << UP::Mark::Nonspacing.new(token, active_opts)
+    when :spacing_mark;           node << UP::Mark::Spacing.new(token, active_opts)
+    when :enclosing_mark;         node << UP::Mark::Enclosing.new(token, active_opts)
+
+    when :number;                 node << UP::Number::Any.new(token, active_opts)
+    when :decimal_number;         node << UP::Number::Decimal.new(token, active_opts)
+    when :letter_number;          node << UP::Number::Letter.new(token, active_opts)
+    when :other_number;           node << UP::Number::Other.new(token, active_opts)
+
+    when :punctuation;            node << UP::Punctuation::Any.new(token, active_opts)
+    when :connector_punctuation;  node << UP::Punctuation::Connector.new(token, active_opts)
+    when :dash_punctuation;       node << UP::Punctuation::Dash.new(token, active_opts)
+    when :open_punctuation;       node << UP::Punctuation::Open.new(token, active_opts)
+    when :close_punctuation;      node << UP::Punctuation::Close.new(token, active_opts)
+    when :initial_punctuation;    node << UP::Punctuation::Initial.new(token, active_opts)
+    when :final_punctuation;      node << UP::Punctuation::Final.new(token, active_opts)
+    when :other_punctuation;      node << UP::Punctuation::Other.new(token, active_opts)
+
+    when :separator;              node << UP::Separator::Any.new(token, active_opts)
+    when :space_separator;        node << UP::Separator::Space.new(token, active_opts)
+    when :line_separator;         node << UP::Separator::Line.new(token, active_opts)
+    when :paragraph_separator;    node << UP::Separator::Paragraph.new(token, active_opts)
+
+    when :symbol;                 node << UP::Symbol::Any.new(token, active_opts)
+    when :math_symbol;            node << UP::Symbol::Math.new(token, active_opts)
+    when :currency_symbol;        node << UP::Symbol::Currency.new(token, active_opts)
+    when :modifier_symbol;        node << UP::Symbol::Modifier.new(token, active_opts)
+    when :other_symbol;           node << UP::Symbol::Other.new(token, active_opts)
+
+    when :other;                  node << UP::Codepoint::Any.new(token, active_opts)
+    when :control;                node << UP::Codepoint::Control.new(token, active_opts)
+    when :format;                 node << UP::Codepoint::Format.new(token, active_opts)
+    when :surrogate;              node << UP::Codepoint::Surrogate.new(token, active_opts)
+    when :private_use;            node << UP::Codepoint::PrivateUse.new(token, active_opts)
+    when :unassigned;             node << UP::Codepoint::Unassigned.new(token, active_opts)
+
+    when *UPTokens::Age;          node << UP::Age.new(token, active_opts)
+    when *UPTokens::Derived;      node << UP::Derived.new(token, active_opts)
+    when *UPTokens::Emoji;        node << UP::Emoji.new(token, active_opts)
+    when *UPTokens::Enumerated;   node << UP::Enumerated.new(token, active_opts)
+    when *UPTokens::Script;       node << UP::Script.new(token, active_opts)
+    when *UPTokens::UnicodeBlock; node << UP::Block.new(token, active_opts)
 
     else
       raise UnknownTokenError.new('UnicodeProperty', token)
@@ -473,86 +477,39 @@ class Regexp::Parser
   end
 
   def quantifier(token)
-    target_node = node.expressions.reverse.find { |exp| !exp.is_a?(FreeSpace) }
-    target_node or raise ParserError, "No valid target found for '#{token.text}'"
+    target_node = node.extract_quantifier_target(token.text)
 
     # in case of chained quantifiers, wrap target in an implicit passive group
     # description of the problem: https://github.com/ammar/regexp_parser/issues/3
     # rationale for this solution: https://github.com/ammar/regexp_parser/pull/69
     if target_node.quantified?
-      new_token = Regexp::Token.new(
-        :group,
-        :passive,
-        '', # text
-        target_node.ts,
-        nil, # te (unused)
-        target_node.level,
-        target_node.set_level,
-        target_node.conditional_level
+      new_group = Group::Passive.construct(
+        token:             :passive,
+        ts:                target_node.ts,
+        level:             target_node.level,
+        set_level:         target_node.set_level,
+        conditional_level: target_node.conditional_level,
+        options:           active_opts,
       )
-      new_group = Group::Passive.new(new_token, active_opts)
       new_group.implicit = true
       new_group << target_node
-      increase_level(target_node)
+      increase_group_level(target_node)
       node.expressions[node.expressions.index(target_node)] = new_group
       target_node = new_group
     end
 
-    case token.token
-    when :zero_or_one
-      target_node.quantify(:zero_or_one, token.text, 0, 1, :greedy)
-    when :zero_or_one_reluctant
-      target_node.quantify(:zero_or_one, token.text, 0, 1, :reluctant)
-    when :zero_or_one_possessive
-      target_node.quantify(:zero_or_one, token.text, 0, 1, :possessive)
-
-    when :zero_or_more
-      target_node.quantify(:zero_or_more, token.text, 0, -1, :greedy)
-    when :zero_or_more_reluctant
-      target_node.quantify(:zero_or_more, token.text, 0, -1, :reluctant)
-    when :zero_or_more_possessive
-      target_node.quantify(:zero_or_more, token.text, 0, -1, :possessive)
-
-    when :one_or_more
-      target_node.quantify(:one_or_more, token.text, 1, -1, :greedy)
-    when :one_or_more_reluctant
-      target_node.quantify(:one_or_more, token.text, 1, -1, :reluctant)
-    when :one_or_more_possessive
-      target_node.quantify(:one_or_more, token.text, 1, -1, :possessive)
-
-    when :interval
-      interval(target_node, token)
-
-    else
+    unless token.token =~ /\A(?:zero_or_one|zero_or_more|one_or_more|interval)
+                             (?:_greedy|_reluctant|_possessive)?\z/x
       raise UnknownTokenError.new('Quantifier', token)
     end
+
+    target_node.quantify(token, active_opts)
   end
 
-  def increase_level(exp)
+  def increase_group_level(exp)
     exp.level += 1
-    exp.respond_to?(:each) && exp.each { |subexp| increase_level(subexp) }
-  end
-
-  def interval(target_node, token)
-    text = token.text
-    mchr = text[text.length-1].chr =~ /[?+]/ ? text[text.length-1].chr : nil
-    case mchr
-    when '?'
-      range_text = text[0...-1]
-      mode = :reluctant
-    when '+'
-      range_text = text[0...-1]
-      mode = :possessive
-    else
-      range_text = text
-      mode = :greedy
-    end
-
-    range = range_text.gsub(/\{|\}/, '').split(',', 2)
-    min = range[0].empty? ? 0 : range[0]
-    max = range[1] ? (range[1].empty? ? -1 : range[1]) : min
-
-    target_node.quantify(:interval, text, min.to_i, max.to_i, mode)
+    exp.quantifier.level += 1 if exp.quantifier
+    exp.terminal? || exp.each { |subexp| increase_group_level(subexp) }
   end
 
   def set(token)
@@ -568,6 +525,8 @@ class Regexp::Parser
   end
 
   def open_set(token)
+    # TODO: this and Quantifier are the only cases where Expression#token
+    # does not match the scanner/lexer output. Fix in v3.0.0.
     token.token = :character
     nest(CharacterSet.new(token, active_opts))
   end
@@ -582,7 +541,7 @@ class Regexp::Parser
 
   def range(token)
     exp = CharacterSet::Range.new(token, active_opts)
-    scope = node.last.is_a?(CharacterSet::IntersectedSequence) ? node.last : node
+    scope = node.last.instance_of?(CharacterSet::IntersectedSequence) ? node.last : node
     exp << scope.expressions.pop
     nest(exp)
   end
@@ -609,26 +568,29 @@ class Regexp::Parser
   end
 
   def close_completed_character_set_range
-    decrease_nesting if node.is_a?(CharacterSet::Range) && node.complete?
+    decrease_nesting if node.instance_of?(CharacterSet::Range) && node.complete?
   end
 
   def active_opts
     options_stack.last
   end
 
-  # Assigns referenced expressions to refering expressions, e.g. if there is
+  # Assigns referenced expressions to referring expressions, e.g. if there is
   # an instance of Backreference::Number, its #referenced_expression is set to
   # the instance of Group::Capture that it refers to via its number.
   def assign_referenced_expressions
-    targets = {}
-    # find all referencable expressions
+    # find all referenceable and referring expressions
+    targets = { 0 => root }
+    referrers = []
     root.each_expression do |exp|
       exp.is_a?(Group::Capture) && targets[exp.identifier] = exp
+      referrers << exp if exp.referential?
     end
-    # assign them to any refering expressions
-    root.each_expression do |exp|
-      exp.respond_to?(:reference) &&
-        exp.referenced_expression = targets[exp.reference]
+    # assign reference expression to referring expressions
+    # (in a second iteration because there might be forward references)
+    referrers.each do |exp|
+      exp.referenced_expression = targets[exp.reference] ||
+        raise(ParserError, "Invalid reference #{exp.reference} at pos #{exp.ts}")
     end
   end
 end # module Regexp::Parser

data/lib/regexp_parser/scanner/errors/premature_end_error.rb

@@ -0,0 +1,8 @@
+class Regexp::Scanner
+  # Unexpected end of pattern
+  class PrematureEndError < ScannerError
+    def initialize(where = '')
+      super "Premature end of pattern at #{where}"
+    end
+  end
+end

data/lib/regexp_parser/scanner/errors/scanner_error.rb

@@ -0,0 +1,6 @@
+require_relative '../../../regexp_parser/error'
+
+class Regexp::Scanner
+  # General scanner error (catch all)
+  class ScannerError < Regexp::Parser::Error; end
+end

data/lib/regexp_parser/scanner/errors/validation_error.rb

@@ -0,0 +1,63 @@
+class Regexp::Scanner
+  # Base for all scanner validation errors
+  class ValidationError < ScannerError
+    # Centralizes and unifies the handling of validation related errors.
+    def self.for(type, problem, reason = nil)
+      types.fetch(type).new(problem, reason)
+    end
+
+    def self.types
+      @types ||= {
+        backref:      InvalidBackrefError,
+        group:        InvalidGroupError,
+        group_option: InvalidGroupOption,
+        posix_class:  UnknownPosixClassError,
+        property:     UnknownUnicodePropertyError,
+        sequence:     InvalidSequenceError,
+      }
+    end
+  end
+
+  # Invalid sequence format. Used for escape sequences, mainly.
+  class InvalidSequenceError < ValidationError
+    def initialize(what = 'sequence', where = '')
+      super "Invalid #{what} at #{where}"
+    end
+  end
+
+  # Invalid group. Used for named groups.
+  class InvalidGroupError < ValidationError
+    def initialize(what, reason)
+      super "Invalid #{what}, #{reason}."
+    end
+  end
+
+  # Invalid groupOption. Used for inline options.
+  # TODO: should become InvalidGroupOptionError in v3.0.0 for consistency
+  class InvalidGroupOption < ValidationError
+    def initialize(option, text)
+      super "Invalid group option #{option} in #{text}"
+    end
+  end
+
+  # Invalid back reference. Used for name a number refs/calls.
+  class InvalidBackrefError < ValidationError
+    def initialize(what, reason)
+      super "Invalid back reference #{what}, #{reason}"
+    end
+  end
+
+  # The property name was not recognized by the scanner.
+  class UnknownUnicodePropertyError < ValidationError
+    def initialize(name, _)
+      super "Unknown unicode character property name #{name}"
+    end
+  end
+
+  # The POSIX class name was not recognized by the scanner.
+  class UnknownPosixClassError < ValidationError
+    def initialize(text, _)
+      super "Unknown POSIX class #{text}"
+    end
+  end
+end