Object#respond_to? and Object#inspect raise exception after allocate

Automatic cleanup of spam.

This issue has been automatically marked as stale because it has not been commented on for at least three months.

The resources of the Rails core team are limited, and so we are asking for your help. If you can still reproduce this error on the 3-0-stable branch or on master, please reply with all of the information you have about it and add "[state:open]" to your comment. This will reopen the ticket for review. Likewise, if you feel that this is a very important feature for Rails to include, please reply with your explanation so we can consider it.

Thank you for all your contributions, and we hope you will understand this step to focus our efforts where they are most helpful.

I can reproduce this. Serialization of objects that rely on #initialize -- particularly AR objects -- fails in AR::Base#object_from_yaml:

      def object_from_yaml(string)

    return string unless string.is_a?(String) && string =~ /^---/
    YAML::load(string) rescue string
  end</code>

Applying the patch attached here (manually, as it doesn't apply as is) fixes the issue completely as far as I can tell (base.rb, starting at line 1505):

      # Returns true if the given attribute is in the attributes hash

  def has_attribute?(attr_name)
    # @attributes.has_key?(attr_name.to_s)
    @attributes.has_key?(attr_name.to_s) if instance_variable_defined?("@attributes")
  end

  # Returns an array of names for the attributes available on this object sorted alphabetically.
  def attribute_names
    # @attributes.keys.sort
    instance_variable_defined?("@attributes") ? @attributes.keys.sort : []
  end</code>

To work around this we tried the wrapping approach suggested by MarianTheisen here: https://rails.lighthouseapp.com/projects/8994/tickets/5525-after-up... to no avail. I actually suspect that bug is only partially fixed and that fixing this one would be the definitive fix for the other as well.

Attaching a barebones rails 3.0.3 app (two models, Beef and Bone; Beef has a serialized column where we want to stick an instance of Bone) where the following console session will demonstrate the breakage:

>> Beef
  SQL (0.6ms)  SHOW TABLES
=> Beef(id: integer, name: string, details: text, created_at: datetime, updated_at: datetime)
>> b = Beef.new(name: 'Rosa', details: [Bone.first])
  Bone Load (0.4ms)  SELECT `bones`.* FROM `bones` LIMIT 1
=> #<Beef id: nil, name: "Rosa", details: [#<Bone id: 1, type: nil, size: 10, created_at: "2011-02-07 09:47:47", updated_at: "2011-02-07 09:47:47">], created_at: nil, updated_at: nil>
>> b.details
=> [#<Bone id: 1, type: nil, size: 10, created_at: "2011-02-07 09:47:47", updated_at: "2011-02-07 09:47:47">]
>> b.save!
  SQL (0.1ms)  BEGIN
  SQL (0.9ms)  describe `beefs`
  AREL (0.2ms)  INSERT INTO `beefs` (`name`, `details`, `created_at`, `updated_at`) VALUES ('Rosa', '--- \n- !ruby/object:Bone \n attributes: \n id: 1\n type: \n size: 10\n created_at: 2011-02-07 09:47:47 Z\n updated_at: 2011-02-07 09:47:47 Z\n attributes_cache: {}\n\n changed_attributes: {}\n\n destroyed: false\n marked_for_destruction: false\n persisted: true\n previously_changed: {}\n\n readonly: false\n', '2011-02-07 09:48:39', '2011-02-07 09:48:39')
  SQL (0.1ms)  COMMIT
=> true
>> b.details
=> [#<Bone id: 1, type: nil, size: 10, created_at: "2011-02-07 09:47:47", updated_at: "2011-02-07 09:47:47">]
>> b.reload.details
  Beef Load (0.3ms)  SELECT `beefs`.* FROM `beefs` WHERE (`beefs`.`id` = 1) LIMIT 1
=> "--- \n- !ruby/object:Bone \n  attributes: \n    id: 1\n    type: \n    size: 10\n    created_at: 2011-02-07 09:47:47 Z\n    updated_at: 2011-02-07 09:47:47 Z\n  attributes_cache: {}\n\n  changed_attributes: {}\n\n  destroyed: false\n  marked_for_destruction: false\n  persisted: true\n  previously_changed: {}\n\n  readonly: false\n"

Please don't hesitate to contact me for further details or information. We would really like to see this bug fixed for 3.0.4

Thank you so much for all your hard work.

David