friendly_id breaks ActiveRecord::Base.belongs_to_required_by_default in Rails 5.0.0

[nflorentin]

In Rails 5.0.0, a new config was introduced to activate deactivate the automatic presence validation on belongs_to associations.

In config/active_record_belongs_to_required_by_default.rb:
Rails.application.config.active_record.belongs_to_required_by_default = true

Without friendly_id in gemfile

> Rails.application.config.active_record.belongs_to_required_by_default
=> true
> ActiveRecord::Base.belongs_to_required_by_default
=> true
> User.belongs_to_required_by_default
=> true

After adding friendly_id in gemfile

In rails console:

> Rails.application.config.active_record.belongs_to_required_by_default
=> true
> ActiveRecord::Base.belongs_to_required_by_default
=> nil
> User.belongs_to_required_by_default
=> nil

[nflorentin]

As a temporary fix, you can replace Rails.application.config.active_record.belongs_to_required_by_default = true with ActiveRecord::Base.belongs_to_required_by_default = true in config/initializers/active_record_belongs_to_required_by_default.rb but I'm not sure if that could produce any side effect.

[kimrgrey]

Hi, @nflorentin! Sounds weird, because, as I know, friendly_id doesn't do anything with this config. It looks like a side effect of something. I'll take a look on this issue.

[arokettu]

@kimrgrey from Rails issue referenced above:

This is due to initialization order. Any plugin that loads AR will cause it to pull config to AR::Base so subsequent initializers will be too late in the game to take effect.

You may close this one

[jprinaldi]

So why was this issue closed? Is there no way FriendlyId can prevent this from happening?

[edwardmp]

I agree with @juampi, that this issue is not resolved and friendly_id should be modified to adapt for this and thus the issue shouldn't be considered closed

[nflorentin]

Here an example of a similar issue addressed for devise heartcombo/devise@c2c74b0

[edwardmp]

@nflorentin great, thanks for linking. Let me try If I can apply this to friendly_id too and submit a PR :)

[jeffblake]

Any luck on this?

[nflorentin]

The issue is still there in version 5.4.2.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.