psych 3.0.3 → 3.3.1

data/ext/psych/yaml/yaml_private.h

@@ -3,7 +3,7 @@
 #endif
 
 #if HAVE_CONFIG_H
-#include <config.h>
+#include "config.h"
 #endif
 
 #include <yaml.h>
@@ -12,16 +12,6 @@
 #include <limits.h>
 #include <stddef.h>
 
-#ifndef _MSC_VER
-#include <stdint.h>
-#else
-#ifdef _WIN64
-#define PTRDIFF_MAX _I64_MAX
-#else
-#define PTRDIFF_MAX INT_MAX
-#endif
-#endif
-
 /*
  * Memory management.
  */
@@ -80,6 +70,17 @@ yaml_parser_fetch_more_tokens(yaml_parser_t *parser);
 
 #define OUTPUT_RAW_BUFFER_SIZE  (OUTPUT_BUFFER_SIZE*2+2)
 
+/*
+ * The maximum size of a YAML input file.
+ * This used to be PTRDIFF_MAX, but that's not entirely portable
+ * because stdint.h isn't available on all platforms.
+ * It is not entirely clear why this isn't the maximum value
+ * that can fit into the parser->offset field.
+ */
+
+#define MAX_FILE_SIZE (~(size_t)0 / 2)
+
+
 /*
  * The size of other stacks and queues.
  */
@@ -93,7 +94,7 @@ yaml_parser_fetch_more_tokens(yaml_parser_t *parser);
  */
 
 #define BUFFER_INIT(context,buffer,size)                                        \
-    (((buffer).start = yaml_malloc(size)) ?                                     \
+  (((buffer).start = (yaml_char_t *)yaml_malloc(size)) ?                        \
         ((buffer).last = (buffer).pointer = (buffer).start,                     \
          (buffer).end = (buffer).start+(size),                                  \
          1) :                                                                   \
@@ -133,7 +134,7 @@ yaml_string_join(
      (value).pointer = (string))
 
 #define STRING_INIT(context,string,size)                                        \
-    (((string).start = yaml_malloc(size)) ?                                     \
+    (((string).start = YAML_MALLOC(size)) ?                                     \
         ((string).pointer = (string).start,                                     \
          (string).end = (string).start+(size),                                  \
          memset((string).start, 0, (size)),                                     \
@@ -174,14 +175,14 @@ yaml_string_join(
  * Check the octet at the specified position.
  */
 
-#define CHECK_AT(string,octet,offset)                                           \
+#define CHECK_AT(string,octet,offset)                   \
     ((string).pointer[offset] == (yaml_char_t)(octet))
 
 /*
  * Check the current octet in the buffer.
  */
 
-#define CHECK(string,octet) CHECK_AT((string),(octet),0)
+#define CHECK(string,octet) (CHECK_AT((string),(octet),0))
 
 /*
  * Check if the character at the specified position is an alphabetical
@@ -423,10 +424,10 @@ yaml_stack_extend(void **start, void **top, void **end);
 YAML_DECLARE(int)
 yaml_queue_extend(void **start, void **head, void **tail, void **end);
 
-#define STACK_INIT(context,stack,size)                                          \
-    (((stack).start = yaml_malloc((size)*sizeof(*(stack).start))) ?             \
+#define STACK_INIT(context,stack,type)                                     \
+  (((stack).start = (type)yaml_malloc(INITIAL_STACK_SIZE*sizeof(*(stack).start))) ? \
         ((stack).top = (stack).start,                                           \
-         (stack).end = (stack).start+(size),                                    \
+         (stack).end = (stack).start+INITIAL_STACK_SIZE,                        \
          1) :                                                                   \
         ((context)->error = YAML_MEMORY_ERROR,                                  \
          0))
@@ -456,8 +457,8 @@ yaml_queue_extend(void **start, void **head, void **tail, void **end);
 #define POP(context,stack)                                                      \
     (*(--(stack).top))
 
-#define QUEUE_INIT(context,queue,size)                                          \
-    (((queue).start = yaml_malloc((size)*sizeof(*(queue).start))) ?             \
+#define QUEUE_INIT(context,queue,size,type)                                     \
+  (((queue).start = (type)yaml_malloc((size)*sizeof(*(queue).start))) ?         \
         ((queue).head = (queue).tail = (queue).start,                           \
          (queue).end = (queue).start+(size),                                    \
          1) :                                                                   \
@@ -660,3 +661,28 @@ yaml_queue_extend(void **start, void **head, void **tail, void **end);
      (node).data.mapping.pairs.end = (node_pairs_end),                          \
      (node).data.mapping.pairs.top = (node_pairs_start),                        \
      (node).data.mapping.style = (node_style))
+
+/* Strict C compiler warning helpers */
+
+#if defined(__clang__) || defined(__GNUC__)
+#  define HASATTRIBUTE_UNUSED
+#endif
+#ifdef HASATTRIBUTE_UNUSED
+#  define __attribute__unused__             __attribute__((__unused__))
+#else
+#  define __attribute__unused__
+#endif
+
+/* Shim arguments are arguments that must be included in your function,
+ * but serve no purpose inside.  Silence compiler warnings. */
+#define SHIM(a) /*@unused@*/ a __attribute__unused__
+
+/* UNUSED_PARAM() marks a shim argument in the body to silence compiler warnings */
+#ifdef __clang__
+#  define UNUSED_PARAM(a) (void)(a);
+#else
+#  define UNUSED_PARAM(a) /*@-noeffect*/if (0) (void)(a)/*@=noeffect*/;
+#endif
+
+#define YAML_MALLOC_STATIC(type) (type*)yaml_malloc(sizeof(type))
+#define YAML_MALLOC(size)        (yaml_char_t *)yaml_malloc(size)

data/lib/psych.rb

@@ -10,11 +10,7 @@ when 'jruby'
     org.jruby.ext.psych.PsychLibrary.new.load(JRuby.runtime, false)
   end
 else
-  begin
-    require "#{RUBY_VERSION[/\d+\.\d+/]}/psych.so"
-  rescue LoadError
-    require 'psych.so'
-  end
+  require 'psych.so'
 end
 require 'psych/nodes'
 require 'psych/streaming'
@@ -36,7 +32,7 @@ require 'psych/class_loader'
 # = Overview
 #
 # Psych is a YAML parser and emitter.
-# Psych leverages libyaml [Home page: http://pyyaml.org/wiki/LibYAML]
+# Psych leverages libyaml [Home page: https://pyyaml.org/wiki/LibYAML]
 # or [HG repo: https://bitbucket.org/xi/libyaml] for its YAML parsing
 # and emitting capabilities. In addition to wrapping libyaml, Psych also
 # knows how to serialize and de-serialize most Ruby objects to and from
@@ -78,12 +74,15 @@ require 'psych/class_loader'
 #
 # ==== Reading from a string
 #
-#   Psych.load("--- a")             # => 'a'
-#   Psych.load("---\n - a\n - b")   # => ['a', 'b']
+#   Psych.safe_load("--- a")             # => 'a'
+#   Psych.safe_load("---\n - a\n - b")   # => ['a', 'b']
+#   # From a trusted string:
+#   Psych.load("--- !ruby/range\nbegin: 0\nend: 42\nexcl: false\n") # => 0..42
 #
 # ==== Reading from a file
 #
-#   Psych.load_file("database.yml")
+#   Psych.safe_load_file("data.yml", permitted_classes: [Date])
+#   Psych.load_file("trusted_database.yml")
 #
 # ==== Exception handling
 #
@@ -234,15 +233,17 @@ require 'psych/class_loader'
 
 module Psych
   # The version of libyaml Psych is using
-  LIBYAML_VERSION = Psych.libyaml_version.join '.'
-
-  FALLBACK        = Struct.new :to_ruby # :nodoc:
+  LIBYAML_VERSION = Psych.libyaml_version.join('.').freeze
+  # Deprecation guard
+  NOT_GIVEN = Object.new.freeze
+  private_constant :NOT_GIVEN
 
   ###
   # Load +yaml+ in to a Ruby data structure.  If multiple documents are
   # provided, the object contained in the first document will be returned.
-  # +filename+ will be used in the exception message if any exception is raised
-  # while parsing.
+  # +filename+ will be used in the exception message if any exception
+  # is raised while parsing.  If +yaml+ is empty, it returns
+  # the specified +fallback+ return value, which defaults to +false+.
   #
   # Raises a Psych::SyntaxError when a YAML syntax error is detected.
   #
@@ -252,7 +253,7 @@ module Psych
   #   Psych.load("---\n - a\n - b")   # => ['a', 'b']
   #
   #   begin
-  #     Psych.load("--- `", "file.txt")
+  #     Psych.load("--- `", filename: "file.txt")
   #   rescue Psych::SyntaxError => ex
   #     ex.file    # => 'file.txt'
   #     ex.message # => "(file.txt): found character that cannot start any token"
@@ -264,11 +265,21 @@ module Psych
   #   Psych.load("---\n foo: bar")                         # => {"foo"=>"bar"}
   #   Psych.load("---\n foo: bar", symbolize_names: true)  # => {:foo=>"bar"}
   #
-  def self.load yaml, filename = nil, fallback: false, symbolize_names: false
-    result = parse(yaml, filename, fallback: fallback)
-    result = result.to_ruby if result
-    symbolize_names!(result) if symbolize_names
-    result
+  # Raises a TypeError when `yaml` parameter is NilClass
+  #
+  # NOTE: This method *should not* be used to parse untrusted documents, such as
+  # YAML documents that are supplied via user input.  Instead, please use the
+  # safe_load method.
+  #
+  def self.load yaml, legacy_filename = NOT_GIVEN, filename: nil, fallback: false, symbolize_names: false, freeze: false
+    if legacy_filename != NOT_GIVEN
+      warn_with_uplevel 'Passing filename with the 2nd argument of Psych.load is deprecated. Use keyword argument like Psych.load(yaml, filename: ...) instead.', uplevel: 1 if $VERBOSE
+      filename = legacy_filename
+    end
+
+    result = parse(yaml, filename: filename)
+    return fallback unless result
+    result.to_ruby(symbolize_names: symbolize_names, freeze: freeze)
   end
 
   ###
@@ -284,27 +295,27 @@ module Psych
   # * Hash
   #
   # Recursive data structures are not allowed by default.  Arbitrary classes
-  # can be allowed by adding those classes to the +whitelist+.  They are
+  # can be allowed by adding those classes to the +permitted_classes+ keyword argument.  They are
   # additive.  For example, to allow Date deserialization:
   #
-  #   Psych.safe_load(yaml, [Date])
+  #   Psych.safe_load(yaml, permitted_classes: [Date])
   #
   # Now the Date class can be loaded in addition to the classes listed above.
   #
-  # Aliases can be explicitly allowed by changing the +aliases+ parameter.
+  # Aliases can be explicitly allowed by changing the +aliases+ keyword argument.
   # For example:
   #
   #   x = []
   #   x << x
   #   yaml = Psych.dump x
   #   Psych.safe_load yaml               # => raises an exception
-  #   Psych.safe_load yaml, [], [], true # => loads the aliases
+  #   Psych.safe_load yaml, aliases: true # => loads the aliases
   #
   # A Psych::DisallowedClass exception will be raised if the yaml contains a
-  # class that isn't in the whitelist.
+  # class that isn't in the +permitted_classes+ list.
   #
   # A Psych::BadAlias exception will be raised if the yaml contains aliases
-  # but the +aliases+ parameter is set to false.
+  # but the +aliases+ keyword argument is set to false.
   #
   # +filename+ will be used in the exception message if any exception is raised
   # while parsing.
@@ -315,20 +326,39 @@ module Psych
   #   Psych.safe_load("---\n foo: bar")                         # => {"foo"=>"bar"}
   #   Psych.safe_load("---\n foo: bar", symbolize_names: true)  # => {:foo=>"bar"}
   #
-  def self.safe_load yaml, whitelist_classes = [], whitelist_symbols = [], aliases = false, filename = nil, symbolize_names: false
-    result = parse(yaml, filename)
-    return unless result
+  def self.safe_load yaml, legacy_permitted_classes = NOT_GIVEN, legacy_permitted_symbols = NOT_GIVEN, legacy_aliases = NOT_GIVEN, legacy_filename = NOT_GIVEN, permitted_classes: [], permitted_symbols: [], aliases: false, filename: nil, fallback: nil, symbolize_names: false, freeze: false
+    if legacy_permitted_classes != NOT_GIVEN
+      warn_with_uplevel 'Passing permitted_classes with the 2nd argument of Psych.safe_load is deprecated. Use keyword argument like Psych.safe_load(yaml, permitted_classes: ...) instead.', uplevel: 1 if $VERBOSE
+      permitted_classes = legacy_permitted_classes
+    end
 
-    class_loader = ClassLoader::Restricted.new(whitelist_classes.map(&:to_s),
-                                               whitelist_symbols.map(&:to_s))
-    scanner      = ScalarScanner.new class_loader
-    if aliases
-      visitor = Visitors::ToRuby.new scanner, class_loader
-    else
-      visitor = Visitors::NoAliasRuby.new scanner, class_loader
+    if legacy_permitted_symbols != NOT_GIVEN
+      warn_with_uplevel 'Passing permitted_symbols with the 3rd argument of Psych.safe_load is deprecated. Use keyword argument like Psych.safe_load(yaml, permitted_symbols: ...) instead.', uplevel: 1 if $VERBOSE
+      permitted_symbols = legacy_permitted_symbols
     end
+
+    if legacy_aliases != NOT_GIVEN
+      warn_with_uplevel 'Passing aliases with the 4th argument of Psych.safe_load is deprecated. Use keyword argument like Psych.safe_load(yaml, aliases: ...) instead.', uplevel: 1 if $VERBOSE
+      aliases = legacy_aliases
+    end
+
+    if legacy_filename != NOT_GIVEN
+      warn_with_uplevel 'Passing filename with the 5th argument of Psych.safe_load is deprecated. Use keyword argument like Psych.safe_load(yaml, filename: ...) instead.', uplevel: 1 if $VERBOSE
+      filename = legacy_filename
+    end
+
+    result = parse(yaml, filename: filename)
+    return fallback unless result
+
+    class_loader = ClassLoader::Restricted.new(permitted_classes.map(&:to_s),
+                                               permitted_symbols.map(&:to_s))
+    scanner      = ScalarScanner.new class_loader
+    visitor = if aliases
+                Visitors::ToRuby.new scanner, class_loader, symbolize_names: symbolize_names, freeze: freeze
+              else
+                Visitors::NoAliasRuby.new scanner, class_loader, symbolize_names: symbolize_names, freeze: freeze
+              end
     result = visitor.accept result
-    symbolize_names!(result) if symbolize_names
     result
   end
 
@@ -344,28 +374,40 @@ module Psych
   #   Psych.parse("---\n - a\n - b") # => #<Psych::Nodes::Document:0x00>
   #
   #   begin
-  #     Psych.parse("--- `", "file.txt")
+  #     Psych.parse("--- `", filename: "file.txt")
   #   rescue Psych::SyntaxError => ex
   #     ex.file    # => 'file.txt'
   #     ex.message # => "(file.txt): found character that cannot start any token"
   #   end
   #
   # See Psych::Nodes for more information about YAML AST.
-  def self.parse yaml, filename = nil, fallback: false
-    parse_stream(yaml, filename) do |node|
+  def self.parse yaml, legacy_filename = NOT_GIVEN, filename: nil, fallback: NOT_GIVEN
+    if legacy_filename != NOT_GIVEN
+      warn_with_uplevel 'Passing filename with the 2nd argument of Psych.parse is deprecated. Use keyword argument like Psych.parse(yaml, filename: ...) instead.', uplevel: 1 if $VERBOSE
+      filename = legacy_filename
+    end
+
+    parse_stream(yaml, filename: filename) do |node|
       return node
     end
-    fallback
+
+    if fallback != NOT_GIVEN
+      warn_with_uplevel 'Passing the `fallback` keyword argument of Psych.parse is deprecated.', uplevel: 1 if $VERBOSE
+      fallback
+    else
+      false
+    end
   end
 
   ###
   # Parse a file at +filename+. Returns the Psych::Nodes::Document.
   #
   # Raises a Psych::SyntaxError when a YAML syntax error is detected.
-  def self.parse_file filename
-    File.open filename, 'r:bom|utf-8' do |f|
-      parse f, filename
+  def self.parse_file filename, fallback: false
+    result = File.open filename, 'r:bom|utf-8' do |f|
+      parse f, filename: filename
     end
+    result || fallback
   end
 
   ###
@@ -394,14 +436,21 @@ module Psych
   #   end
   #
   #   begin
-  #     Psych.parse_stream("--- `", "file.txt")
+  #     Psych.parse_stream("--- `", filename: "file.txt")
   #   rescue Psych::SyntaxError => ex
   #     ex.file    # => 'file.txt'
   #     ex.message # => "(file.txt): found character that cannot start any token"
   #   end
   #
+  # Raises a TypeError when NilClass is passed.
+  #
   # See Psych::Nodes for more information about YAML AST.
-  def self.parse_stream yaml, filename = nil, &block
+  def self.parse_stream yaml, legacy_filename = NOT_GIVEN, filename: nil, &block
+    if legacy_filename != NOT_GIVEN
+      warn_with_uplevel 'Passing filename with the 2nd argument of Psych.parse_stream is deprecated. Use keyword argument like Psych.parse_stream(yaml, filename: ...) instead.', uplevel: 1 if $VERBOSE
+      filename = legacy_filename
+    end
+
     if block_given?
       parser = Psych::Parser.new(Handlers::DocumentStream.new(&block))
       parser.parse yaml, filename
@@ -502,68 +551,113 @@ module Psych
   #   end
   #   list # => ['foo', 'bar']
   #
-  def self.load_stream yaml, filename = nil
-    if block_given?
-      parse_stream(yaml, filename) do |node|
-        yield node.to_ruby
-      end
-    else
-      parse_stream(yaml, filename).children.map { |child| child.to_ruby }
+  def self.load_stream yaml, legacy_filename = NOT_GIVEN, filename: nil, fallback: [], **kwargs
+    if legacy_filename != NOT_GIVEN
+      warn_with_uplevel 'Passing filename with the 2nd argument of Psych.load_stream is deprecated. Use keyword argument like Psych.load_stream(yaml, filename: ...) instead.', uplevel: 1 if $VERBOSE
+      filename = legacy_filename
     end
+
+    result = if block_given?
+               parse_stream(yaml, filename: filename) do |node|
+                 yield node.to_ruby(**kwargs)
+               end
+             else
+               parse_stream(yaml, filename: filename).children.map { |node| node.to_ruby(**kwargs) }
+             end
+
+    return fallback if result.is_a?(Array) && result.empty?
+    result
   end
 
   ###
   # Load the document contained in +filename+.  Returns the yaml contained in
   # +filename+ as a Ruby object, or if the file is empty, it returns
   # the specified +fallback+ return value, which defaults to +false+.
-  def self.load_file filename, fallback: false
+  #
+  # NOTE: This method *should not* be used to parse untrusted documents, such as
+  # YAML documents that are supplied via user input.  Instead, please use the
+  # safe_load_file method.
+  def self.load_file filename, **kwargs
+    File.open(filename, 'r:bom|utf-8') { |f|
+      self.load f, filename: filename, **kwargs
+    }
+  end
+
+  ###
+  # Safely loads the document contained in +filename+.  Returns the yaml contained in
+  # +filename+ as a Ruby object, or if the file is empty, it returns
+  # the specified +fallback+ return value, which defaults to +false+.
+  # See safe_load for options.
+  def self.safe_load_file filename, **kwargs
     File.open(filename, 'r:bom|utf-8') { |f|
-      self.load f, filename, fallback: FALLBACK.new(fallback)
+      self.safe_load f, filename: filename, **kwargs
     }
   end
 
   # :stopdoc:
-  @domain_types = {}
   def self.add_domain_type domain, type_tag, &block
     key = ['tag', domain, type_tag].join ':'
-    @domain_types[key] = [key, block]
-    @domain_types["tag:#{type_tag}"] = [key, block]
+    domain_types[key] = [key, block]
+    domain_types["tag:#{type_tag}"] = [key, block]
   end
 
   def self.add_builtin_type type_tag, &block
     domain = 'yaml.org,2002'
     key = ['tag', domain, type_tag].join ':'
-    @domain_types[key] = [key, block]
+    domain_types[key] = [key, block]
   end
 
   def self.remove_type type_tag
-    @domain_types.delete type_tag
+    domain_types.delete type_tag
   end
 
-  @load_tags = {}
-  @dump_tags = {}
   def self.add_tag tag, klass
-    @load_tags[tag] = klass.name
-    @dump_tags[klass] = tag
+    load_tags[tag] = klass.name
+    dump_tags[klass] = tag
   end
 
-  def self.symbolize_names!(result)
-    case result
-    when Hash
-      result.keys.each do |key|
-        result[key.to_sym] = symbolize_names!(result.delete(key))
-      end
-    when Array
-      result.map! { |r| symbolize_names!(r) }
+  # Workaround for emulating `warn '...', uplevel: 1` in Ruby 2.4 or lower.
+  def self.warn_with_uplevel(message, uplevel: 1)
+    at = parse_caller(caller[uplevel]).join(':')
+    warn "#{at}: #{message}"
+  end
+
+  def self.parse_caller(at)
+    if /^(.+?):(\d+)(?::in `.*')?/ =~ at
+      file = $1
+      line = $2.to_i
+      [file, line]
     end
-    result
   end
-  private_class_method :symbolize_names!
+  private_class_method :warn_with_uplevel, :parse_caller
 
   class << self
-    attr_accessor :load_tags
-    attr_accessor :dump_tags
-    attr_accessor :domain_types
+    if defined?(Ractor)
+      require 'forwardable'
+      extend Forwardable
+
+      class Config
+        attr_accessor :load_tags, :dump_tags, :domain_types
+        def initialize
+          @load_tags = {}
+          @dump_tags = {}
+          @domain_types = {}
+        end
+      end
+
+      def config
+        Ractor.current[:PsychConfig] ||= Config.new
+      end
+
+      def_delegators :config, :load_tags, :dump_tags, :domain_types, :load_tags=, :dump_tags=, :domain_types=
+    else
+      attr_accessor :load_tags
+      attr_accessor :dump_tags
+      attr_accessor :domain_types
+    end
   end
+  self.load_tags = {}
+  self.dump_tags = {}
+  self.domain_types = {}
   # :startdoc:
 end