Rails Ruby on Rails

Comments and changes to this ticket

• You flagged this item as spam.
  

  Hates_ January 8th, 2010 @ 11:46 AM

  • Tag changed from “accepts_nested_attributes_for, activerecord” to “accepts_nested_attributes_for, activerecord, fixed”

  Tested with 2-3-stable branch and this seems to have been fixed.

• 

  Eloy Duran January 11th, 2010 @ 10:18 AM

  • State changed from “new” to “resolved”
  • Assigned user set to “Eloy Duran”

  I was just gonna ask that :) Thanks for the update.

• You flagged this item as spam.
  

  Alessio April 6th, 2010 @ 03:53 PM

  Hello Richard,

  I have the same problem than you! I am with Rails 2.3.5, how have you solved this problem? Rails 2.3.5 isn't a 2-3-stable branch? (I am begginer)

  Thank you!

  Alessio

• You flagged this item as spam.
  

  Yossi April 7th, 2010 @ 11:13 PM

  Eloy,
  I have the same problem (Rails 2.3.5)
  Can you tell me how did you resolved it?

  Thanks,
  Yossi

• 

  Eloy Duran April 8th, 2010 @ 09:42 AM

  Rails 2.3.5 was released at the end of november last year: http://weblog.rubyonrails.org/2009/11/30/ruby-on-rails-2-3-5-released. So seeing as Richard confirmed it was fixed on 2-3-stable in january this year, it means the fix is not in a release yet.

  Until 2.3.6 is released, I'd advice you to do a checkout of http://github.com/rails/rails/tree/2-3-stable.

  HTH

• You flagged this item as spam.
  

  Liam Morley (carpeliam) June 26th, 2010 @ 05:09 PM

  • Importance changed from “” to “Low”

  Which commit was this, or which public release did this get rolled in? As I saw this in 2.3.5, upgraded to 2.3.8, and I'm still seeing it. Here is my model:

  class User < ActiveRecord::Base
    has_many :conferences, :through => :conference_attendees, :order => 'end_date DESC'
    has_many :conference_attendees
    accepts_nested_attributes_for, :conferences, :allow_destroy => true
  end
  
  class Conference < ActiveRecord::Base
    has_many :users, :through => :conference_attendees
    has_many :conference_attendees, :dependent => :destroy
    accepts_nested_attributes_for :conference_attendees, :allow_destroy => true
  end
  
  class ConferenceAttendee < ActiveRecord::Base
    belongs_to :conference
    belongs_to :user
  end
  
  @user = User.first
  # @user.id == 2
  params = {"user"=>{"conferences_attributes"=>{"0"=>{"name"=>"Conf Name", "conference_attendees_attributes"=>{"0"=>{"role"=>"Attendee", "user_id"=>"1984"}}}}}}
  @user.update_attributes params['user']
  

  When I run this in Rails 2.3.8, I see the following in my logs:
  

    Conference Create (0.2ms)   INSERT INTO tech_conferences (name, created_at, updated_at) VALUES('Conf Name', '2010-06-26 03:01:07', '2010-06-26 03:01:07')
    ConferenceAttendee Create (0.2ms)   INSERT INTO tech_conference_attendees (created_at, updated_at, role, user_id, conference_id) VALUES('2010-06-26 03:01:07', '2010-06-26 03:01:07', 'Attendee', 1984, 1124)
    ConferenceAttendee Create (0.7ms)   INSERT INTO tech_conference_attendees (created_at, updated_at, role, user_id, conference_id) VALUES('2010-06-26 03:01:07', '2010-06-26 03:01:07', NULL, 2, 1124)
  

  This varies slightly from the original poster. In my case, the second conference has a NULL value for the 'role' field in ConferenceAttendee, and also has the same user id as the @user although that ID was not specified in the params for ConferenceAttendee.

  Also of note: I tried changing my accepts_nested_attributes_for :conference_attendees and added a :reject_if that rejects everything, and ran this again. The first ConferenceAttendee is not created, but the second one is.

  My quick-and-dirty workaround for this right now is an after_save that deletes the record if it doesn't have a 'role' attribute, I'm just hoping this kluge doesn't get me fired. :) Please, if you've got a better temporary solution, I'm all ears.

• You flagged this item as spam.
  

  Satish Chauhan July 7th, 2010 @ 12:33 PM

  I have the same problem (Rails 2.3.5)

  Its resolved in 2.3.8.

  any idea about the patch that resolved this problem?