Reader / predicate returning different values for boolean field

[thiagogsr]

[33] pry(main)> UserTask.find(13511).task.counselor
  UserTask Load (1.9ms)  SELECT  "user_tasks".* FROM "user_tasks"  WHERE "user_tasks"."id" = $1 LIMIT 1  [["id", 13511]]
  TaskItem Load (0.7ms)  SELECT  "task_items".* FROM "task_items"  WHERE "task_items"."id" = $1 LIMIT 1  [["id", 241]]
=> false
[34] pry(main)> UserTask.find(13511).task.counselor?
  UserTask Load (0.9ms)  SELECT  "user_tasks".* FROM "user_tasks"  WHERE "user_tasks"."id" = $1 LIMIT 1  [["id", 13511]]
  TaskItem Load (1.0ms)  SELECT  "task_items".* FROM "task_items"  WHERE "task_items"."id" = $1 LIMIT 1  [["id", 241]]
=> true

[6] pry(main)> UserTask.find(13511).task.id
  UserTask Load (1.0ms)  SELECT  "user_tasks".* FROM "user_tasks"  WHERE "user_tasks"."id" = $1 LIMIT 1  [["id", 13511]]
  TaskItem Load (1.0ms)  SELECT  "task_items".* FROM "task_items"  WHERE "task_items"."id" = $1 LIMIT 1  [["id", 241]]
=> 241
[7] pry(main)> TaskItem.find(241).counselor?
  TaskItem Load (1.1ms)  SELECT  "task_items".* FROM "task_items"  WHERE "task_items"."id" = $1 LIMIT 1  [["id", 241]]
=> true
[8] pry(main)> TaskItem.find(241).counselor
  TaskItem Load (0.9ms)  SELECT  "task_items".* FROM "task_items"  WHERE "task_items"."id" = $1 LIMIT 1  [["id", 241]]
=> false

• The value on database is false
• The field is boolean
• Rails version: 4.1.6
• Ruby version: ruby 2.1.3p242 (2014-09-19 revision 47630) [x86_64-linux]
• PostgreSQL: 9.3

Is it a bug?

[thiagogsr]

rails/activesupport/test/core_ext/class/attribute_test.rb

Line 30 in 08754f1

test 'predicate method' do

This method should be so:

test 'predicate method' do
    assert_equal false, @klass.setting?
    @klass.setting = 1
    assert_equal true, @klass.setting?
  end

or

test 'predicate method' do
    @klass.setting = 0
    assert_equal false, @klass.setting?
    @klass.setting = 1
    assert_equal true, @klass.setting?
  end

On first way is @klass.setting false or nil?

[sgrif]

This is the expected behavior. The predicate checks for emptiness, not truthiness. Please see the documentation. http://api.rubyonrails.org/classes/ActiveRecord/AttributeMethods.html#method-i-attribute_present-3F

[thiagogsr]

Yep @sgrif, but others fields have other behaviors

[21] pry(main)> TaskItem.find(241).ancestor
  TaskItem Load (1.1ms)  SELECT  "task_items".* FROM "task_items"  WHERE "task_items"."id" = $1 LIMIT 1  [["id", 241]]
=> false
[22] pry(main)> TaskItem.find(241).ancestor?
  TaskItem Load (0.8ms)  SELECT  "task_items".* FROM "task_items"  WHERE "task_items"."id" = $1 LIMIT 1  [["id", 241]]
=> false

• The value on database is false
• The field is boolean

[sgrif]

Your pry output is making it difficult to understand your problem. Can you please provide an executable test case using this as a template? https://github.com/rails/rails/blob/master/guides/bug_report_templates/active_record_master.rb

[rafaelfranca]

Something very weird is going on. The expected behaviour is to return false for false values. See https://github.com/rails/rails/blob/master/activerecord/lib/active_record/attribute_methods/query.rb#L15.

[sgrif]

Whoops, you are correct I got the suffix mixed up.

[kuldeepaggarwal]

Everything is working fine, here is the gist.

Ruby version: ruby 2.1.4p265 (2014-10-27 revision 48166) [x86_64-darwin13.0]
PostgreSQL: 9.3.2

[thiagogsr]

@kuldeepaggarwal This problem happened in one field, the reason I don't know. Other boolean fields works!!

[guilhermeFranco]

This could be happening because you may have an enum that defines a value with the same name as the boolean field in your table.

Since enum values results in predicate methods, you're getting the false value because your object doesn't have that enum "type" assigned to it. The question is: In case we have a field with the same name as a enum value, what should be the correct behavior ? Ask the database field or the enum ?

[thiagogsr]

@guilhermeFranco is right. I miss the enum declaration with the same name as the boolean field. BTW, shouldn't the FW inform us? enum task_type: [ :junior, :counselor, :senior ]

[rafaelfranca]

@thiagogsr it does it but for methods already defined by the framework. User defined methods don't give any warning.

[thiagogsr]

@rafaelfranca I asked because the FW raises a fatal error when I try to use a enum item with same name of any relationship

example:

class User < ActiveRecord::Base
  enum role: [ :admin, :parent, :manager ]
  belongs_to :parent
end

[thiagogsr]

Test with the problem

[kuldeepaggarwal]

@thiagogsr You have a boolean field as well for counselor, so rails will create 2 methods counselor?, one by boolean field and one by enum.

[kuldeepaggarwal]

I think Rails behaviour is correct in this case.

[thiagogsr]

@kuldeepaggarwal yep, I just suggested an 'stack level too deep' on this case

[prathamesh-sonpatki]

How about prefixing enum generated methods like enum_parent?

[guilhermeFranco]

@prathamesh-sonpatki I would prefer a combination of the enum name and its value. Given the following model:

class User < ActiveRecord::Base
  enum role: [ :admin, :parent, :manager ]
  belongs_to :parent
end

The enum value verification could be:

user.role_admin?

[prathamesh-sonpatki]

Awesome. This makes sense 😄

[prathamesh-sonpatki]

Something like this? prathamesh-sonpatki@c314864

[kuldeepaggarwal]

I think then all methods should be prefixed when created by enum.

[kuldeepaggarwal]

I think we should raise error or give warning in this scenario.

[rails-bot]

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 team are limited, and so we are asking for your help.

If you can still reproduce this error on the 4-2-stable, 4-1-stable branches or on master,
please reply with all of the information you have about it in order to keep the issue open.

Thank you for all your contributions.

[kuldeepaggarwal]

We can close this issue as per the discussion in #17506

[kamipo]

Closing this since #19813 has already been merged.