The model_name method should be called on to_model
#18647
Conversation
|
Actually I think we should call set |
|
Is there a reason it doesn't happen immediately when doing the https://github.com/rails/rails/blob/v4.2.0/actionview/lib/action_view/helpers/tags/base.rb#L41 |
|
I don't think so. Try it and see how tests responds to the change. |
|
Yeah, never mind, they blew up. I'll try to find that out for myself later. I'll just add the Thanks for the help :) |
mcls
force-pushed
the
placeholderable-to-model
branch
from
b688687 to
505447f
Compare
|
Hmm... found very similar code here: https://github.com/rails/rails/blob/v4.2.0/actionview/lib/action_view/helpers/tags/label.rb#L20-L32 I'll try to clean that up |
mcls
force-pushed
the
placeholderable-to-model
branch
from
505447f to
c946b2a
Compare
|
Not too sure about that last commit. It reduces the duplication, but it could probably be done in a clearer way. |
The `model_name` method should be called on `to_model`
…AR form_for Refactoring at rails#18647 broke using non active record objects in form_for. This patch restores the original behaviour where we only compute i18 key when object.respond_to?(:to_model)
|
There is a bug in the merged version: @model = object.respond_to?(:to_model) ? object.to_model : objectwe should not set form_for(OpenStuct.new(....),)see #18834 |
…AR form_for Refactoring at rails#18647 broke using non active record objects in form_for. This patch restores the original behaviour where we only compute i18 key when object.respond_to?(:to_model)
See #18646
Should the same be done for the following lines? E.g.
model.respond_to?(:human_attribute_name)https://github.com/rails/rails/blob/v4.2.0/actionview/lib/action_view/helpers/tags/placeholderable.rb#L22-L24