psych 3.1.0 → 5.2.3

data/ext/psych/psych_to_ruby.c

@@ -36,4 +36,3 @@ void Init_psych_to_ruby(void)
     rb_define_private_method(cPsychVisitorsToRuby, "build_exception", build_exception, 2);
     rb_define_private_method(class_loader, "path2class", path2class, 1);
 }
-/* vim: set noet sws=4 sw=4: */

data/ext/psych/psych_yaml_tree.c

@@ -2,23 +2,10 @@
 
 VALUE cPsychVisitorsYamlTree;
 
-/*
- * call-seq: private_iv_get(target, prop)
- *
- * Get the private instance variable +prop+ from +target+
- */
-static VALUE private_iv_get(VALUE self, VALUE target, VALUE prop)
-{
-    return rb_attr_get(target, rb_intern(StringValueCStr(prop)));
-}
-
 void Init_psych_yaml_tree(void)
 {
     VALUE psych     = rb_define_module("Psych");
     VALUE visitors  = rb_define_module_under(psych, "Visitors");
     VALUE visitor   = rb_define_class_under(visitors, "Visitor", rb_cObject);
     cPsychVisitorsYamlTree = rb_define_class_under(visitors, "YAMLTree", visitor);
-
-    rb_define_private_method(cPsychVisitorsYamlTree, "private_iv_get", private_iv_get, 2);
 }
-/* vim: set noet sws=4 sw=4: */

data/lib/psych/class_loader.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
-require 'psych/omap'
-require 'psych/set'
+require_relative 'omap'
+require_relative 'set'
 
 module Psych
   class ClassLoader # :nodoc:
@@ -35,9 +35,11 @@ module Psych
 
     constants.each do |const|
       konst = const_get const
-      define_method(const.to_s.downcase) do
-        load konst
-      end
+      class_eval <<~RUBY, __FILE__, __LINE__ + 1
+        def #{const.to_s.downcase}
+          load #{konst.inspect}
+        end
+      RUBY
     end
 
     private
@@ -69,7 +71,7 @@ module Psych
       rescue
         nil
       end
-    }.compact]
+    }.compact].freeze
 
     class Restricted < ClassLoader
       def initialize classes, symbols
@@ -84,7 +86,7 @@ module Psych
         if @symbols.include? sym
           super
         else
-          raise DisallowedClass, 'Symbol'
+          raise DisallowedClass.new('load', 'Symbol')
         end
       end
 
@@ -94,7 +96,7 @@ module Psych
         if @classes.include? klassname
           super
         else
-          raise DisallowedClass, klassname
+          raise DisallowedClass.new('load', klassname)
         end
       end
     end

data/lib/psych/core_ext.rb

@@ -15,5 +15,5 @@ class Object
 end
 
 if defined?(::IRB)
-  require 'psych/y'
+  require_relative 'y'
 end

data/lib/psych/exception.rb

@@ -6,9 +6,23 @@ module Psych
   class BadAlias < Exception
   end
 
+  # Subclasses `BadAlias` for backwards compatibility
+  class AliasesNotEnabled < BadAlias
+    def initialize
+      super "Alias parsing was not enabled. To enable it, pass `aliases: true` to `Psych::load` or `Psych::safe_load`."
+    end
+  end
+
+  # Subclasses `BadAlias` for backwards compatibility
+  class AnchorNotDefined < BadAlias
+    def initialize anchor_name
+      super "An alias referenced an unknown anchor: #{anchor_name}"
+    end
+  end
+
   class DisallowedClass < Exception
-    def initialize klass_name
-      super "Tried to load unspecified class: #{klass_name}"
+    def initialize action, klass_name
+      super "Tried to #{action} unspecified class: #{klass_name}"
     end
   end
 end

data/lib/psych/handler.rb

@@ -119,7 +119,7 @@ module Psych
     # +tag+ is an associated tag or nil
     # +plain+ is a boolean value
     # +quoted+ is a boolean value
-    # +style+ is an integer idicating the string style
+    # +style+ is an integer indicating the string style
     #
     # See the constants in Psych::Nodes::Scalar for the possible values of
     # +style+

data/lib/psych/handlers/document_stream.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
-require 'psych/tree_builder'
+require_relative '../tree_builder'
 
 module Psych
   module Handlers

data/lib/psych/handlers/recorder.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
-require 'psych/handler'
+require_relative '../handler'
 
 module Psych
   module Handlers

data/lib/psych/json/stream.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
-require 'psych/json/ruby_events'
-require 'psych/json/yaml_events'
+require_relative 'ruby_events'
+require_relative 'yaml_events'
 
 module Psych
   module JSON

data/lib/psych/json/tree_builder.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
-require 'psych/json/yaml_events'
+require_relative 'yaml_events'
 
 module Psych
   module JSON

data/lib/psych/nodes/node.rb

@@ -1,7 +1,6 @@
 # frozen_string_literal: true
-require 'stringio'
-require 'psych/class_loader'
-require 'psych/scalar_scanner'
+require_relative '../class_loader'
+require_relative '../scalar_scanner'
 
 module Psych
   module Nodes
@@ -46,8 +45,8 @@ module Psych
       # Convert this node to Ruby.
       #
       # See also Psych::Visitors::ToRuby
-      def to_ruby
-        Visitors::ToRuby.create.accept(self)
+      def to_ruby(symbolize_names: false, freeze: false, strict_integer: false)
+        Visitors::ToRuby.create(symbolize_names: symbolize_names, freeze: freeze, strict_integer: strict_integer).accept(self)
       end
       alias :transform :to_ruby
 
@@ -56,6 +55,7 @@ module Psych
       #
       # See also Psych::Visitors::Emitter
       def yaml io = nil, options = {}
+        require "stringio"
         real_io = io || StringIO.new(''.encode('utf-8'))
 
         Visitors::Emitter.new(real_io, options).accept self

data/lib/psych/nodes/scalar.rb

@@ -50,7 +50,7 @@ module Psych
       # +tag+ is an associated tag or nil
       # +plain+ is a boolean value
       # +quoted+ is a boolean value
-      # +style+ is an integer idicating the string style
+      # +style+ is an integer indicating the string style
       #
       # == See Also
       #

data/lib/psych/nodes.rb

@@ -1,11 +1,11 @@
 # frozen_string_literal: true
-require 'psych/nodes/node'
-require 'psych/nodes/stream'
-require 'psych/nodes/document'
-require 'psych/nodes/sequence'
-require 'psych/nodes/scalar'
-require 'psych/nodes/mapping'
-require 'psych/nodes/alias'
+require_relative 'nodes/node'
+require_relative 'nodes/stream'
+require_relative 'nodes/document'
+require_relative 'nodes/sequence'
+require_relative 'nodes/scalar'
+require_relative 'nodes/mapping'
+require_relative 'nodes/alias'
 
 module Psych
   ###

data/lib/psych/parser.rb

@@ -48,5 +48,18 @@ module Psych
       @handler = handler
       @external_encoding = ANY
     end
+
+    ###
+    # call-seq:
+    #    parser.parse(yaml)
+    #
+    # Parse the YAML document contained in +yaml+.  Events will be called on
+    # the handler set on the parser instance.
+    #
+    # See Psych::Parser and Psych::Parser#handler
+
+    def parse yaml, path = yaml.respond_to?(:path) ? yaml.path : "<unknown>"
+      _native_parse @handler, yaml, path
+    end
   end
 end

data/lib/psych/scalar_scanner.rb

@@ -1,5 +1,4 @@
 # frozen_string_literal: true
-require 'strscan'
 
 module Psych
   ###
@@ -9,103 +8,97 @@ module Psych
     TIME = /^-?\d{4}-\d{1,2}-\d{1,2}(?:[Tt]|\s+)\d{1,2}:\d\d:\d\d(?:\.\d*)?(?:\s*(?:Z|[-+]\d{1,2}:?(?:\d\d)?))?$/
 
     # Taken from http://yaml.org/type/float.html
-    FLOAT = /^(?:[-+]?([0-9][0-9_,]*)?\.[0-9]*([eE][-+][0-9]+)?(?# base 10)
-              |[-+]?\.(inf|Inf|INF)(?# infinity)
-              |\.(nan|NaN|NAN)(?# not a number))$/x
-
-    # Taken from http://yaml.org/type/int.html
-    INTEGER = /^(?:[-+]?0b[0-1_]+          (?# base 2)
-                  |[-+]?0[0-7_]+           (?# base 8)
-                  |[-+]?(?:0|[1-9][0-9_]*) (?# base 10)
-                  |[-+]?0x[0-9a-fA-F_]+    (?# base 16))$/x
+    # Base 60, [-+]inf and NaN are handled separately
+    FLOAT = /^(?:[-+]?([0-9][0-9_,]*)?\.[0-9]*([eE][-+][0-9]+)?(?# base 10))$/x
+
+    # Taken from http://yaml.org/type/int.html and modified to ensure at least one numerical symbol exists
+    INTEGER_STRICT = /^(?:[-+]?0b[_]*[0-1][0-1_]*             (?# base 2)
+                         |[-+]?0[_]*[0-7][0-7_]*              (?# base 8)
+                         |[-+]?(0|[1-9][0-9_]*)               (?# base 10)
+                         |[-+]?0x[_]*[0-9a-fA-F][0-9a-fA-F_]* (?# base 16))$/x
+
+    # Same as above, but allows commas.
+    # Not to YML spec, but kept for backwards compatibility
+    INTEGER_LEGACY = /^(?:[-+]?0b[_,]*[0-1][0-1_,]*                (?# base 2)
+                         |[-+]?0[_,]*[0-7][0-7_,]*                 (?# base 8)
+                         |[-+]?(?:0|[1-9](?:[0-9]|,[0-9]|_[0-9])*) (?# base 10)
+                         |[-+]?0x[_,]*[0-9a-fA-F][0-9a-fA-F_,]*    (?# base 16))$/x
 
     attr_reader :class_loader
 
     # Create a new scanner
-    def initialize class_loader
-      @string_cache = {}
+    def initialize class_loader, strict_integer: false
       @symbol_cache = {}
       @class_loader = class_loader
+      @strict_integer = strict_integer
     end
 
     # Tokenize +string+ returning the Ruby object
     def tokenize string
       return nil if string.empty?
-      return string if @string_cache.key?(string)
       return @symbol_cache[string] if @symbol_cache.key?(string)
-
-      case string
+      integer_regex = @strict_integer ? INTEGER_STRICT : INTEGER_LEGACY
       # Check for a String type, being careful not to get caught by hash keys, hex values, and
       # special floats (e.g., -.inf).
-      when /^[^\d\.:-]?[A-Za-z_\s!@#\$%\^&\*\(\)\{\}\<\>\|\/\\~;=]+/, /\n/
-        if string.length > 5
-          @string_cache[string] = true
-          return string
-        end
+      if string.match?(%r{^[^\d.:-]?[[:alpha:]_\s!@#$%\^&*(){}<>|/\\~;=]+}) || string.match?(/\n/)
+        return string if string.length > 5
 
-        case string
-        when /^[^ytonf~]/i
-          @string_cache[string] = true
+        if string.match?(/^[^ytonf~]/i)
           string
-        when '~', /^null$/i
+        elsif string == '~' || string.match?(/^null$/i)
           nil
-        when /^(yes|true|on)$/i
+        elsif string.match?(/^(yes|true|on)$/i)
           true
-        when /^(no|false|off)$/i
+        elsif string.match?(/^(no|false|off)$/i)
           false
         else
-          @string_cache[string] = true
           string
         end
-      when TIME
+      elsif string.match?(TIME)
         begin
           parse_time string
         rescue ArgumentError
           string
         end
-      when /^\d{4}-(?:1[012]|0\d|\d)-(?:[12]\d|3[01]|0\d|\d)$/
-        require 'date'
+      elsif string.match?(/^\d{4}-(?:1[012]|0\d|\d)-(?:[12]\d|3[01]|0\d|\d)$/)
         begin
-          class_loader.date.strptime(string, '%Y-%m-%d')
+          class_loader.date.strptime(string, '%F', Date::GREGORIAN)
         rescue ArgumentError
           string
         end
-      when /^\.inf$/i
+      elsif string.match?(/^\+?\.inf$/i)
         Float::INFINITY
-      when /^-\.inf$/i
+      elsif string.match?(/^-\.inf$/i)
         -Float::INFINITY
-      when /^\.nan$/i
+      elsif string.match?(/^\.nan$/i)
         Float::NAN
-      when /^:./
+      elsif string.match?(/^:./)
         if string =~ /^:(["'])(.*)\1/
           @symbol_cache[string] = class_loader.symbolize($2.sub(/^:/, ''))
         else
           @symbol_cache[string] = class_loader.symbolize(string.sub(/^:/, ''))
         end
-      when /^[-+]?[0-9][0-9_]*(:[0-5]?[0-9]){1,2}$/
+      elsif string.match?(/^[-+]?[0-9][0-9_]*(:[0-5]?[0-9]){1,2}$/)
         i = 0
         string.split(':').each_with_index do |n,e|
           i += (n.to_i * 60 ** (e - 2).abs)
         end
         i
-      when /^[-+]?[0-9][0-9_]*(:[0-5]?[0-9]){1,2}\.[0-9_]*$/
+      elsif string.match?(/^[-+]?[0-9][0-9_]*(:[0-5]?[0-9]){1,2}\.[0-9_]*$/)
         i = 0
         string.split(':').each_with_index do |n,e|
           i += (n.to_f * 60 ** (e - 2).abs)
         end
         i
-      when FLOAT
-        if string =~ /\A[-+]?\.\Z/
-          @string_cache[string] = true
+      elsif string.match?(FLOAT)
+        if string.match?(/\A[-+]?\.\Z/)
           string
         else
-          Float(string.gsub(/[,_]|\.([Ee]|$)/, '\1'))
+          Float(string.delete(',_').gsub(/\.([Ee]|$)/, '\1'))
         end
+      elsif string.match?(integer_regex)
+        parse_int string
       else
-        int = parse_int string.gsub(/[,_]/, '')
-        return int if int
-
-        @string_cache[string] = true
         string
       end
     end
@@ -113,8 +106,7 @@ module Psych
     ###
     # Parse and return an int from +string+
     def parse_int string
-      return unless INTEGER === string
-      Integer(string)
+      Integer(string.delete(',_'))
     end
 
     ###

data/lib/psych/syntax_error.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
-require 'psych/exception'
+require_relative 'exception'
 
 module Psych
   class SyntaxError < Psych::Exception

data/lib/psych/tree_builder.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
-require 'psych/handler'
+require_relative 'handler'
 
 module Psych
   ###
@@ -41,7 +41,7 @@ module Psych
       Sequence
       Mapping
     }.each do |node|
-      class_eval %{
+      class_eval <<~RUBY, __FILE__, __LINE__ + 1
         def start_#{node.downcase}(anchor, tag, implicit, style)
           n = Nodes::#{node}.new(anchor, tag, implicit, style)
           set_start_location(n)
@@ -54,7 +54,7 @@ module Psych
           set_end_location(n)
           n
         end
-      }
+        RUBY
     end
 
     ###

data/lib/psych/versions.rb

@@ -1,10 +1,10 @@
-
 # frozen_string_literal: true
+
 module Psych
   # The version of Psych you are using
-  VERSION = '3.1.0' unless defined?(::Psych::VERSION)
+  VERSION = '5.2.3'
 
   if RUBY_ENGINE == 'jruby'
-    DEFAULT_SNAKEYAML_VERSION = '1.23'.freeze
+    DEFAULT_SNAKEYAML_VERSION = '2.9'.freeze
   end
 end

data/lib/psych/visitors/json_tree.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
-require 'psych/json/ruby_events'
+require_relative '../json/ruby_events'
 
 module Psych
   module Visitors

data/lib/psych/visitors/to_ruby.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
-require 'psych/scalar_scanner'
-require 'psych/class_loader'
-require 'psych/exception'
+require_relative '../scalar_scanner'
+require_relative '../class_loader'
+require_relative '../exception'
 
 unless defined?(Regexp::NOENCODING)
   Regexp::NOENCODING = 32
@@ -12,39 +12,44 @@ module Psych
     ###
     # This class walks a YAML AST, converting each node to Ruby
     class ToRuby < Psych::Visitors::Visitor
-      def self.create
+      def self.create(symbolize_names: false, freeze: false, strict_integer: false)
         class_loader = ClassLoader.new
-        scanner      = ScalarScanner.new class_loader
-        new(scanner, class_loader)
+        scanner      = ScalarScanner.new class_loader, strict_integer: strict_integer
+        new(scanner, class_loader, symbolize_names: symbolize_names, freeze: freeze)
       end
 
       attr_reader :class_loader
 
-      def initialize ss, class_loader
+      def initialize ss, class_loader, symbolize_names: false, freeze: false
         super()
         @st = {}
         @ss = ss
+        @load_tags = Psych.load_tags
         @domain_types = Psych.domain_types
         @class_loader = class_loader
+        @symbolize_names = symbolize_names
+        @freeze = freeze
       end
 
       def accept target
         result = super
-        return result if @domain_types.empty? || !target.tag
 
-        key = target.tag.sub(/^[!\/]*/, '').sub(/(,\d+)\//, '\1:')
-        key = "tag:#{key}" unless key =~ /^(?:tag:|x-private)/
+        unless @domain_types.empty? || !target.tag
+          key = target.tag.sub(/^[!\/]*/, '').sub(/(,\d+)\//, '\1:')
+          key = "tag:#{key}" unless key.match?(/^(?:tag:|x-private)/)
 
-        if @domain_types.key? key
-          value, block = @domain_types[key]
-          return block.call value, result
+          if @domain_types.key? key
+            value, block = @domain_types[key]
+            result = block.call value, result
+          end
         end
 
+        result = deduplicate(result).freeze if @freeze
         result
       end
 
       def deserialize o
-        if klass = resolve_class(Psych.load_tags[o.tag])
+        if klass = resolve_class(@load_tags[o.tag])
           instance = klass.allocate
 
           if instance.respond_to?(:init_with)
@@ -74,8 +79,9 @@ module Psych
           class_loader.big_decimal._load o.value
         when "!ruby/object:DateTime"
           class_loader.date_time
-          require 'date' unless defined? DateTime
-          @ss.parse_time(o.value).to_datetime
+          t = @ss.parse_time(o.value)
+          DateTime.civil(*t.to_a[0, 6].reverse, Rational(t.utc_offset, 86400)) +
+            (t.subsec/86400)
         when '!ruby/encoding'
           ::Encoding.find o.value
         when "!ruby/object:Complex"
@@ -94,7 +100,7 @@ module Psych
           source  = $1
           options = 0
           lang    = nil
-          ($2 || '').split('').each do |option|
+          $2&.each_char do |option|
             case option
             when 'x' then options |= Regexp::EXTENDED
             when 'i' then options |= Regexp::IGNORECASE
@@ -124,7 +130,7 @@ module Psych
       end
 
       def visit_Psych_Nodes_Sequence o
-        if klass = resolve_class(Psych.load_tags[o.tag])
+        if klass = resolve_class(@load_tags[o.tag])
           instance = klass.allocate
 
           if instance.respond_to?(:init_with)
@@ -156,8 +162,8 @@ module Psych
       end
 
       def visit_Psych_Nodes_Mapping o
-        if Psych.load_tags[o.tag]
-          return revive(resolve_class(Psych.load_tags[o.tag]), o)
+        if @load_tags[o.tag]
+          return revive(resolve_class(@load_tags[o.tag]), o)
         end
         return revive_hash(register(o, {}), o) unless o.tag
 
@@ -252,6 +258,8 @@ module Psych
 
           e = build_exception((resolve_class($1) || class_loader.exception),
                               h.delete('message'))
+
+          e.set_backtrace h.delete('backtrace') if h.key? 'backtrace'
           init_with(e, h, o)
 
         when '!set', 'tag:yaml.org,2002:set'
@@ -316,10 +324,11 @@ module Psych
       end
 
       def visit_Psych_Nodes_Alias o
-        @st.fetch(o.anchor) { raise BadAlias, "Unknown alias: #{o.anchor}" }
+        @st.fetch(o.anchor) { raise AnchorNotDefined, o.anchor }
       end
 
       private
+
       def register node, object
         @st[node.anchor] = object if node.anchor
         object
@@ -331,13 +340,12 @@ module Psych
         list
       end
 
-      SHOVEL = '<<'
-      def revive_hash hash, o
+      def revive_hash hash, o, tagged= false
         o.children.each_slice(2) { |k,v|
           key = accept(k)
           val = accept(v)
 
-          if key == SHOVEL && k.tag != "tag:yaml.org,2002:str"
+          if key == '<<' && k.tag != "tag:yaml.org,2002:str"
             case v
             when Nodes::Alias, Nodes::Mapping
               begin
@@ -359,6 +367,12 @@ module Psych
               hash[key] = val
             end
           else
+            if !tagged && @symbolize_names && key.is_a?(String)
+              key = key.to_sym
+            elsif !@freeze
+              key = deduplicate(key)
+            end
+
             hash[key] = val
           end
 
@@ -366,12 +380,32 @@ module Psych
         hash
       end
 
+      if RUBY_VERSION < '2.7'
+        def deduplicate key
+          if key.is_a?(String)
+            # It is important to untaint the string, otherwise it won't
+            # be deduplicated into an fstring, but simply frozen.
+            -(key.untaint)
+          else
+            key
+          end
+        end
+      else
+        def deduplicate key
+          if key.is_a?(String)
+            -key
+          else
+            key
+          end
+        end
+      end
+
       def merge_key hash, key, val
       end
 
       def revive klass, node
         s = register(node, klass.allocate)
-        init_with(s, revive_hash({}, node), node)
+        init_with(s, revive_hash({}, node, true), node)
       end
 
       def init_with o, h, node
@@ -394,7 +428,7 @@ module Psych
 
     class NoAliasRuby < ToRuby
       def visit_Psych_Nodes_Alias o
-        raise BadAlias, "Unknown alias: #{o.anchor}"
+        raise AliasesNotEnabled
       end
     end
   end